Tokushima University / Educator and Researcher Directory --- Ikeda, Kenji

Personal Web Page

http://www-b2.is.tokushima-u.ac.jp/~ikeda

https://researchmap.jp/read0017422 (researchmap)

Field of Study

○	Control Theory

Subject of Study

○	Autonomous decentralized adaptive control (adaptive control, decentralized control)

Book / Paper

Book:

1.	Kenji Ikeda : システム同定, オーム社, Sep. 2013. (Summary) 7編制御・システム第4章システム同定(pp.357--363)を担当. システム同定の手順, システム同定モデル, ノンパラメトリック同定, パラメトリック同定，部分空間同定法，閉ループ同定について概説した．システム同定の基礎的事項の簡単なまとめと，実際に使用する際に問題となる事項を簡潔に述べている．
2.	新誠一, Kenji Ikeda, 湯浅秀男 and 藤田博之 : 自律分散システム, 朝倉書店, Tokyo, Oct. 1995. (Summary) 自律分散システムの原理と個別のシステムにおける具体的な方法論を提供している. 第1章:自律分散システムの基礎概念, 第2章:ソフトコンピューティング, 第3章:自律分散適応制御, 第4章:自律分散原理による画像認識, 第5章:自律分散とマイクロマシンという構成で, 自律分散原理を基礎に, センシング, 演算, アクチュエータという個別の機能, そしてそれらの統括するシステム論からなってい

Academic Paper (Judged Full Paper):

1.	Michiru Katayama, Kenji Ikeda and Tetsushi Ueta : Stabilization of laminars in chaos intermittency, International Journal of Bifurcation and Chaos, Vol.34, No.2, 2450024-1-2450024-14, 2024. (Summary) Chaos intermittency is composed of a laminar regime, which exhibits almost periodic motion, and a burst regime, which exhibits chaotic motion; it is known that in chaos intermittency, switching between these regimes occurs irregularly. In the laminar regime of chaos intermittency, the periodic solution before the saddle node bifurcation is closely related to its generation, and its behavior becomes periodic in a short time; the laminar is not, however, a periodic solution, and there are no unstable periodic solutions nearby. Most chaos control methods cannot be applied to the problem of stabilizing a laminar response to a periodic solution since they refer to information about unstable periodic orbits. In this paper, we demonstrate a control method that can be applied to the control target that laminar phase of a dynamical system exhibiting chaos intermittency. This method records the time series of a periodic solution prior to the saddle node bifurcation as a pseudo-periodic orbit and feeds it to the control target back. We report that when this control method is applied to a circuit model, laminar motion can be stabilized to a periodic solution via control inputs of very small magnitude, and robust control can be obtained. (Keyword) chaos control / chaos intermittency / bifurcation / OGY method / laminar regime (Tokushima University Institutional Repository) ● Metadata: 118840 (Link to Publication Site) ● Publication site (DOI): 10.1142/S021812742450024X (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85188031320 (Tokushima University Institutional Repository: 118840, DOI: 10.1142/S021812742450024X, Elsevier: Scopus)
2.	Tomoaki Kashiwao, Mayu Hiura, Kenji Ikeda, Mikio Deguchi and Bahadori Alireza : Investigation of effects of phosphor particles on optimal design of surface-mount-device light-emitting diode packaging using ray-tracing simulation, IET Optoelectronics, Vol.12, No.2, 86-93, 2018. (Summary) This study presents investigation of effects of phosphor particles on optimal structure and optical design of surfacemount-device (SMD) light-emitting diode (LED) packaging (PKG) using ray-tracing simulation. In the authors'previous study, effects of the reflection properties of the PKG resin and the lead-frame silver plating on the light extraction efficiency was investigated, and the optimisation method of the cavity angle composing the light reflector of the PKG was proposed. In this study, the optimal cavity angle and reflection properties under the condition of phosphor particles present in mould resin of the top-view PKG are investigated. In addition, the better conditions of the reflection properties under the conditions in which the phosphor is contained and not contained in the mould resin are confirmed. Finally, an optical design method of the SMD LED PKG with the mould resin containing a phosphor is proposed on the basis of the simulation results. (Link to Publication Site) ● Publication site (DOI): 10.1049/iet-opt.2017.0051 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.1049/iet-opt.2017.0051 (DOI: 10.1049/iet-opt.2017.0051)
3.	Tomoaki Kashiwao, Koichi Nakayama, Shin Ando, Kenji Ikeda, Moonyong Lee and Alireza Bahadori : A neural network-based local rainfall prediction system using meteorological data on the Internet, --- A case study using data from the Japan Meteorological Agency ---, Applied Soft Computing, Vol.56, 317-330, 2017. (Summary) In this study, we develop and test a local rainfall (precipitation) prediction system based on arti cial neural networks (ANNs). Our system can automatically obtain meteorological data used for rainfall prediction from the Internet. Meteorological data from equipment installed at a local point is also shared among users in our system. The final goal of the study was the practical use of ``big data'' on the Internet as well as the sharing of data among users for accurate rainfall prediction. We predicted local rainfall in regions of Japan using data from the Japan Meteorological Agency (JMA). As neural network (NN) models for the system, we used a multi-layer perceptron (MLP) with a hybrid algorithm composed of back-propagation (BP) and random optimization (RO) methods, and radial basis function network (RBFN) with a least squares method (LSM), and compared the prediction performance of the two models. Precipitation (total amount of rainfall above 0.5 mm between 12:00 and 24:00 JST (Japan standard time))at Matsuyama, Sapporo, and Naha in 2012 was predicted by NNs using meteorological data for each city from 2011. The volume of precipitation was also predicted (total amount above 1.0 mm between 17:00 and 24:00 JST) at 16 points in Japan and compared with predictions by the JMA in order to verify the universality of the proposed system. The experimental results showed that precipitation in Japan can be predicted by the proposed method, and that the prediction performance of the MLP model was superior to that of the RBFN model for the rainfall prediction problem. However, the results were not better than those generated by the JMA. Finally, heavy rainfall (above 10 mm/h) in summer (Jun. Sep.) afternoons(12:00 24:00 JST) in Tokyo in 2011 and 2012 was predicted using data for Tokyo between 2000 and 2010.The results showed that the volume of precipitation could be accurately predicted and the caching rate of heavy rainfall was high. This suggests that the proposed system can predict unexpected local heavy rainfalls as ``guerrilla rainstorms." (Keyword) local rainfall prediction / precipitation / meteorological data / neural networks / big data (Link to Publication Site) ● Publication site (DOI): 10.1016/j.asoc.2017.03.015 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.1016/j.asoc.2017.03.015 (DOI: 10.1016/j.asoc.2017.03.015)
4.	Tomoaki Kashiwao, Mayu Hiura, Yan Yee Lim, Alireza Bahadori, Kenji Ikeda and Mikio Deguchi : Optimization of surface-mount-device light-emitting diode packaging, --- investigation of effects of component optical properties on light extraction efficiency ---, Optical Engineering, Vol.55, No.2, 025101-1-025101-11, 2016. (Summary) An investigation of the effects of the optical properties of surface-mount-device (SMD) light-emitting diode (LED) (side-view and top-view LEDs) packaging (PKG) components on the light extraction efficiency ηPKG using ray-tracing simulations is presented. In particular, it is found that the optical properties of the PKG resin and the lead-frame (L/F) silver-plating significantly affect ηPKG. Thus, the effects of the surface reflection methods of these components are investigated in order to optimize the optical design of the LED PKG. It is shown that there exists peak extraction efficiency for each PKG, and the cavity angle formed by the cavity wall is important to the optical design. In addition, the effect of phosphor present in the mold resin is examined using a Mie scattering simulation. Finally, an SMD LED PKG optical design method is proposed on the basis of the simulation results. (Tokushima University Institutional Repository) ● Metadata: 115676 (Link to Publication Site) ● Publication site (DOI): 10.1117/1.OE.55.2.025101 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.1117/1.OE.55.2.025101 (Tokushima University Institutional Repository: 115676, DOI: 10.1117/1.OE.55.2.025101)
5.	Takao Shimomura, Kenji Ikeda and Montanaro Adriano : Distribution and Sharing of E-Learning Courseware by Weighted Keywords, International Journal of Information and Education Technology (IJIET), Vol.5, No.9, 698-703, 2014. (Summary) Web-based training is popular around the world, where a large number of people can study at the same time through the Internet anytime and anywhere. The major issue of current e-learning is courseware distribution between different platforms. SCORM (sharable content object reference model) is one of the solutions for courseware sharing. However, to make SCORM-conformable courseware, some knowledge about HTML and JavaScript is required. We have developed a SWF (Sharable Web Fragment)-based e-learning system, where courseware is created with sharable Web fragments such as Web pages, images and other resources, and the courseware can be distributed to another platform by export and import facilities. The exported meta-information is architecture-independent and provides a model of courseware distribution. This paper presents how to promote the distribution and sharing of Web-based e-learning courseware, where courseware can be easily retrieved by weighted keywords. (Link to Publication Site) ● Publication site (DOI): 10.7763/IJIET.2015.V5.595 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.7763/IJIET.2015.V5.595 (DOI: 10.7763/IJIET.2015.V5.595)
6.	Takao Shimomura and Kenji Ikeda : Two Types of Deadlock Detection: Cyclic and Acyclic, Intelligent Systems for Science and Information, Studies in Computational Intelligence, Vol.542, 233-259, 2014. (Summary) Concurrent programs are difficult to test and debug due to their non-deterministic execution. For deadlocks, traditional deadlock detection algorithms depend on finding cycles in lock graphs created from application programs. This paper first introduces three kinds of blocked relations, lock-blocked, wait-blocked, and join-blocked for Java multi-threaded programs. Previous work does not consider the wait-blocked relations, nor the influence of thread interruption. The paper then proposes two types of deadlocks based on these blocked relations, that is, block-cycle type deadlocks and waiting-block type deadlocks which are acyclic. It also presents an example of implementation to detect these types of deadlocks, and addresses future directions. (Link to Publication Site) ● Publication site (DOI): 10.1007/978-3-319-04702-7_14 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-84958530569 (DOI: 10.1007/978-3-319-04702-7_14, Elsevier: Scopus)
7.	Tomoaki Kashiwao, Kenji Ikeda and Takao Shimomura : Estimation of Transition Times of a Hopping Machine Based on Wavelet Analysis, SICE Journal of Control, Measurement, and System Integration, Vol.6, No.6, 396-402, 2013. (Summary) The transition time of a hopping machine (HM) as a piecewise linear (PWL) system is estimated by using a wavelet analysis method proposed in previous papers. The HM is a nonlinear system because its equation of motion has a nonlinear drag coefficient term owing to air resistance. The effectiveness of the proposed method for real (nonlinear) systems is verified through instrumental experiments performed on the HM. First, the effect of nonlinearity on the estimation of a transition time is examined. The HM has a discrete transition caused by the nonlinearity in a real system. In addition, two types of discrete transitions are detected from the output data: one caused by the change between discrete free-fall and spring-mass states and the other caused by inversion of the sign of the nonlinear term owing to a change in the velocity of the mass. These transition times are estimated using the proposed method. Finally, the proposed method is compared with a conventional method based on a clustering technique. (Keyword) piecewise linear system / transition time / wavelet analysis / hopping machine / nonlinearity (Link to Publication Site) ● Publication site (DOI): 10.9746/jcmsi.6.396 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.9746/jcmsi.6.396 (DOI: 10.9746/jcmsi.6.396)
8.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : BCLS Method for Stationary Systems in Closed Loop Environment, SICE Journal of Control, Measurement, and System Integration, Vol.5, No.2, 78-84, 2012. (Summary) This paper proposes a bias-compensated least squares (BCLS) method in closed loop environment when the noise is assumed to be stationary. The feedback controller is assumed to be linear time invariant discrete-time system and the feedback system is assumed to be asymptotically stable. The convergence of the proposed algorithm is analyzed and it is shown that a kind of persistent excitation (PE) condition on the reference input is required. (Keyword) parameter estimation / least squares method / closed loop systems / bias compensation (Link to Publication Site) ● Publication site (DOI): 10.9746/jcmsi.5.78 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.9746/jcmsi.5.78 (DOI: 10.9746/jcmsi.5.78)
9.	Takao Shimomura, Kenji Ikeda and Takahashi Muneo : Functional Web Component Generation for Visual Web Application Programming, International Journal of Computers and Applications, Vol.33, No.2, 167-174, 2011. (Link to Publication Site) ● Publication site (DOI): 10.2316/Journal.202.2011.2.202-2730 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.2316/Journal.202.2011.2.202-2730 (DOI: 10.2316/Journal.202.2011.2.202-2730)
10.	Min Zheng, Kenji Ikeda and Takao Shimomura : Estimation of Continuous-Time Nonlinear Systems by Using the Unscented Kalman Filter, SICE Journal of Control, Measurement, and System Integration, Vol.3, No.5, 324-329, 2010. (Summary) This paper proposes a continuous-time model estimation method by using the Unscented Kalman Filter (UKF) from sampled I/O data, in which plant parameters as well as the initial state are estimated. The initial state is estimated based on a backward system of the plant, and the parameters are estimated by using an iterated UKF, which repeats the estimation of the forward system and the backward system alternately. In order to demonstrate the effectiveness of the proposed method, a rotary pendulum is considered to estimate the parameters of a continuous-time nonlinear system. (Keyword) nonlinear systems / system idenfitication and estimation / control application (Link to Publication Site) ● Publication site (DOI): 10.9746/jcmsi.3.324 (Link to Search Site for Scientific Articles) ● CiNii @ National Institute of Informatics (CRID): 1360846646334119552 ● Search Scopus @ Elsevier (DOI): 10.9746/jcmsi.3.324 (DOI: 10.9746/jcmsi.3.324, CiNii: 1360846646334119552)
11.	Tomoaki Kashiwao, Kenji Ikeda and Takao Shimomura : Estimation of the transition time of piecewise linear systems by using wavelet analysis, Transactions of the Society of Instrument and Control Engineers, Vol.45, No.7, 349-355, 2009. (Summary) This paper proposes an estimation method of the transition time of hybrid systems in piecewise linear (PWL) form by using wavelet analysis. If each transition time is known, the continuous dynamics between adjacent discrete transitions will be identified by conventional identification method for continuous systems. The discrete transition is represented as a switching of differential equations or a jump of continous states. These changes can be detected as characteristic patterns in wavelet transform of an output signal. First, detectability of the discrete transition is defined and its necessary and sufficient condition is derived. Under the assumption of the detectability of th ediscrete transition, an output signal is represented by using Taylor expansion around the transition time. In addition, the characteristic patterns in the wavelet transform of an output signal are analyzed. An estimation method of the transition time based on the analyzed pattern is proposed. Finally, the effectiveness of the proposed method is illustrated by numerical examples. (Keyword) hybrid systems / piecewise linear systems / transition time / wavelet analysis (Link to Publication Site) ● Publication site (DOI): 10.9746/sicetr.45.349 (Link to Search Site for Scientific Articles) ● CiNii @ National Institute of Informatics (CRID): 1390001204504071424 ● Search Scopus @ Elsevier (DOI): 10.9746/sicetr.45.349 (DOI: 10.9746/sicetr.45.349, CiNii: 1390001204504071424)
12.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias Compensation of Recursive Least Squares Estimate in Closed Loop Environment, SICE Journal of Control, Measurement, and System Integration, Vol.1, No.5, 400-405, 2008. (Summary) In this paper, an asymptotic bias of the recursive least squares (RLS) estimate in the closed loop environment is analyzed and its compensation method is proposed under the assumption that the noise is white. Namely, a bias compensated RLS method in the closed loop environment based on output error (OE) model is proposed. A posteriori error is also analyzed for the estimation of the noise variance. (Keyword) recursive least squares estimate / bias compensation / closed loop identification (Link to Publication Site) ● Publication site (DOI): 10.9746/jcmsi.1.400 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.9746/jcmsi.1.400 (DOI: 10.9746/jcmsi.1.400)
13.	Takao Shimomura, Lang Sok Nhor, Chen Liang Quan and Kenji Ikeda : Context-Dependent Extensible Syntax-Oriented Verifier with Recursive Verification, WSEAS Transactions on Information Science and Applications, Vol.5, No.2, 44-53, 2008. (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-42149099519 (Elsevier: Scopus)
14.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-compensated Least Squares Method in Closed Loop Environment, Transactions of the Society of Instrument and Control Engineers, Vol.43, No.12, 1097-1102, 2007. (Summary) In this paper, a bias-compensated least squares method in the closed loop environment is proposed. It is assumed that the observation noise is a white gaussian signal while there are no process noises. It is also assumed that the plant is controlled by a linear time invariant controller and that the closed loop system is asymptotically stable. It is shown that the asymptotic bias of the least squares estimate in the closed loop environment is represented as a difference between the parameter of the prefilter and that of an "optimal prefilter" and that the proposed bias-compensated least squares estimate is consistent. (Keyword) parameter estimation / least squares method / closed loop system / bias compensation (Link to Search Site for Scientific Articles) ● CiNii @ National Institute of Informatics (CRID): 1520853832639592448 (CiNii: 1520853832639592448)
15.	Takao Shimomura, Chen Liang Quan, Lang Sok Nhor and Kenji Ikeda : Web-based Laboratory Network Management System that Integrates Multiple Servers, WSEAS Transactions on Information Science and Applications, Vol.4, No.12, 1422-1430, 2007. (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-41149166085 (Elsevier: Scopus)
16.	Takao Shimomura, Chen Liang Quan, Lang Sok Nhor and Kenji Ikeda : Uniform Tag-based Rich Component Generation for Web Application Development, WSEAS Transactions on Information Science and Applications, Vol.4, No.12, 1431-1438, 2007. (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-41149129173 (Elsevier: Scopus)
17.	Takao Shimomura, Lang Sok Nhor, Chen Liang Quan and Kenji Ikeda : Customizable Pattern-Oriented Verifiers forWeb Applications, WSEAS Transactions on Information Science and Applications, Vol.4, No.12, 1439-1446, 2007. (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-41149152177 (Elsevier: Scopus)
18.	Takao Shimomura, Kenji Ikeda, Chen Liang Quan, Lang Sok Nhor and Takahashi Muneo : Visual Programming of Hierarchical Frames for Web Applications, WSEAS Transactions on Information Science and Applications, Vol.4, No.5, 968-975, 2007. (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-34248326038 (Elsevier: Scopus)
19.	Takao Shimomura, Takahashi Muneo, Chen Liang Quan, Lang Sok Nhor and Kenji Ikeda : Easy enjoyable effective automated practice and test system, INTERNATIONAL JOURNAL of EDUCATION AND INFORMATION TECHNOLOGIES, Vol.1, No.1, 1-7, 2007.
20.	Takanori Tagami, Kenji Ikeda and Tohru Kawabe : Design of Robust I-PD Controller by Using Pareto-Optimality, Transactions of the Institute of Systems, Control and Information Engineers, Vol.17, No.9, 363-370, 2004. (Summary) 先の論文では, 速応性と振動抑制性の2目的最適化問題を, 閉ループ系の極の実部の最大値の最小化, および, 原点から見た虚軸からの角度の最大値の最小化という2つの最小化問題に置き換えて, 2目的問題を解いた. 一方, 筆者の一人は出力誤差の2乗積分の最小化, および, 出力誤差の微分値の2乗積分の最小化という最適化問題を解く方法を提案している. そこで, 本論文では, 二つの手法の特徴を実際の磁気浮上系を例題にとり, 比較·検討した. (Keyword) Pareto-optimality / robust pole assignment / genetic algorithm (Link to Publication Site) ● Publication site (DOI): 10.5687/iscie.17.363 (Link to Search Site for Scientific Articles) ● CiNii @ National Institute of Informatics (CRID): 1390001205164812928 ● Search Scopus @ Elsevier (DOI): 10.5687/iscie.17.363 (DOI: 10.5687/iscie.17.363, CiNii: 1390001205164812928)
21.	Takao Shimomura, Takahashi Muneo, Kenji Ikeda and Yoshio Mogami : Visual Design for Server-Side Programs and Program Generation, Transactions of Information Processing Society of Japan, Vol.45, No.7, 1737-1744, 2004. (Summary) Webアプリケーションの開発において，開発者が頭の中で描いたイメージに従ってソフトウェアを開発することを可能とする，グラフィックスを用いたイメージ指向プログラミング技法を提案し，データベースやプログラムテーブルの内容に基づいて，Webテーブルを動的に構成するWebアプリケーション自動合成方式について述べる．Webコンポーネント間の関係の一貫性，完全性をユーザが容易に検証でき，設計の途中でも，ユーザがそれまでに設計したWebコンポーネントの実行時の様子を容易に確認できる．
22.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-time Model Identification by Using Adaptive Observer, Transactions of the Society of Instrument and Control Engineers, Vol.40, No.4, 415-423, 2004. (Summary) This paper proposes a new method of continuous-time model identification from sampled I/O data by using an adaptive observer. The boundedness of the parameter estimate and the exponential convergence of the parameter estimate error to 0 under the PE assumption are guaranteed. In order to identify the plant from a finite number of I/O data, an adaptive observer of a backward system is also proposed. (Keyword) continous-time model / identification / adaptive observer / sampled-data system (Link to Search Site for Scientific Articles) ● CiNii @ National Institute of Informatics (CRID): 1520572357091461760 (CiNii: 1520572357091461760)
23.	Takanori Tagami and Kenji Ikeda : Design of robust pole assignment based on Pareto-optimal solutions, Asian Journal of Control, Vol.5, No.2, 195-205, 2003. (Summary) PID型の補償器を用いた制御系設計においては, 速応性と振動抑制性のトレードオフをとりながら, パラメータを調整することが多い. この調整は一般に試行錯誤によって行われ, 経験を積んだ現場のオペレータでないと調整は難しいと言われている. 筆者らは, この調整を速応性と振動抑制性の2目的最適化問題として捉え, 遺伝的アルゴリズムによる多点探索を用い, パレート最適解集合をオペレータに提示することによって, 調整における試行錯誤の低減を図った. (Keyword) robust pole assignment / polytope of polynomials / edge theorem / genetic algorithm / Pareto-optimal solutions (Link to Publication Site) ● Publication site (DOI): 10.1111/j.1934-6093.2003.tb00111.x (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-0141928910 (DOI: 10.1111/j.1934-6093.2003.tb00111.x, Elsevier: Scopus)
24.	Takao Shimomura, Muneo Takahashi, Kenji Ikeda and Yoshio Mogami : Web application generator by image-oriented design, ACM SIGSOFT Software Engineering Notes, Vol.28, No.2, 14-19, 2003. (Summary) Webアプリケーションの開発において，開発者が頭の中で描いたイメージに従ってソフトウェアを開発することを可能とする，グラフィックスを用いたイメージ指向プログラミング技法を提案する．本論分では，ビジュアルなWebコンポーネントやWebページの遷移に基づいて実行されるビジネスロジックを組み合わせて，Webアプリケーションプログラムを合成するためのアプリケーション自動生成方式について述べる．
25.	Kenji Ikeda and Seiichi Shin : Fault Tolerance of Autonomous Decentralized Adaptive Control Systems, International Journal of Systems Science, Vol.29, No.7, 773-782, 1998. (Summary) 分散型制御系の特徴として, 耐故障性, 保守·点検の容易さ, オンライン拡張性などが挙げられている. 本論文では, 先の論文で提案した分散型ハイゲイン適応制御が耐故障性を持つことを, 自己相似性に基づいて示し, 分散型システムが耐故障性をもつための原則の一つを明らかにした.
26.	Kenji Ikeda and Seiichi Shin : Adaptive stabilization of minimum phase plants with relative degree `n` or `n+1`, Transactions of the Institute of Systems, Control and Information Engineers, Vol.11, No.6, 290-296, 1998. (Summary) 先の論文で, 耐故障性を有する分散型適応制御系の構成において, 拡張誤差信号を用いたモデルマッチング制御は有効ではなく, ハイゲイン適応制御が有効であることを示した. しかしながら, ハイゲイン適応制御の適用に当たっては, 各サブシステムの相対次数が1であるという制約を必要とした. この制約を緩和するため, もう一つの正実化手法である backstepping 法を用いた分散型適応制御系を構成し, その安定性を証明した.
27.	Kenji Ikeda and Shin Seiichi : A design of decentralized high gain adapvie control system using backstepping, Transactions of the Society of Instrument and Control Engineers, Vol.32, No.10, 1399-1406, 1996. (Summary) 先の論文で, 耐故障性を有する分散型適応制御系の構成において, 拡張誤差信号を用いたモデルマッチング制御は有効ではなく, ハイゲイン適応制御が有効であることを示した. しかしながら, ハイゲイン適応制御の適用に当たっては, 各サブシステムの相対次数が1であるという制約を必要とした. この制約を緩和するため, もう一つの正実化手法である backstepping 法を用いた分散型適応制御系を構成し, その安定性を証明した. (Keyword) decentralize control / high gain control / adaptive control / fault tolerance / backstepping
28.	Takahara Shigeyuki, Kenji Ikeda and Miyamoto Sadaaki : Fuzzy Control Rules and Stability Conditions, Proc. of the 2nd International Conference on Fuzzy Logic and Neural Networks, 425-428, 1992. (Summary) ファジィ制御システムの安定性をナイキスト安定判別条件と絶対安定性条件の2点から考察した. 積和法とミニマックス法を取り上げ, 積和法の方が安定性を保証するパラメータの領域が広いという意味で, ミニマックス法に比べて優れているこをと示した.
29.	Kenji Ikeda, Shin Seiichi and Kitamori Toshiyuki : A design of decentralized high gain adaptive control systems, Transactions of the Society of Instrument and Control Engineers, Vol.28, No.5, 555-563, 1992. (Summary) いくつかのサブシステムが相互に結合した分散型システムを, 各サブシステムの情報のみに基づいて適応的に安定化する, 分散型ハイゲイン適応制御系を提案し, その安定性を証明した. 各サブシステムは最小位相系で相対次数が1, サブシステム間の干渉はマッチング条件を満たすという仮定のもと, シグマ修正則を導入したハイゲイン適応制御によって, 全状態が有界となることを示した. また, 一部のサブシステムの補償器が故障して制御入力が0になっても全システムの安定性が保たれる, 耐故障性についても解析した. (Keyword) decentralized systems / high gain control / adaptive control / stability
30.	Kenji Ikeda, Shin Seiichi and Kitamori Toshiyuki : Robustness of MRACS for linear plants with unknown relative degrees, Transactions of the Society of Instrument and Control Engineers, Vol.27, No.6, 671-677, 1991. (Summary) 先に提案した相対次数に依存しない連続時間モデル規範型適応制御系は, 外乱が存在しない場合には出力誤差の0への収束性を示すことができたが, 外乱や観測雑音が存在する場合は収束しない. そこで, 外乱や観測雑音に対するロバスト化手法の一つである不感帯を用いたパラメータ調整則を導入して, パラメータの収束性, 内部状態および出力誤差の有界性を示した. (Keyword) robust control / adaptive control / relative degree / stability / model reference control
31.	Kenji Ikeda and Kitamori Toshiyuki : A design of MRACS for linear plants with unknown relative degrees, Transactions of the Society of Instrument and Control Engineers, Vol.27, No.1, 49-55, 1991. (Summary) 制御対象の相対次数は入力から出力までの位相遅れの最大値を表す指標である. モデル規範型適応制御においてパラメータ調整を行う場合, 入力から出力までの位相遅れを補正して, 線形部を正実化しなければ安定性を保証することができない. 拡張誤差信号を用いる手法に代表される正実化手法は, 基本的には制御対象の相対次数という事前情報が必要であった. 本論文では, その制約を緩和して, 相対次数の上限がわかっているという仮定の下, モデル規範型適応制御系の提案と, パラメータの有界性と出力誤差の0への収束性を示した. (Keyword) adaptive control / relative degree / stability / model reference control

Academic Paper (Unrefereed Paper):

1.	Takao Shimomura, Kenji Ikeda and Yoshio Mogami : Automatic Program Generation based on Visual Image-Oriented Design, Bulletin of Faculty of Engineering, The University of Tokushima, Vol.48, No.48, 5, 2003. (Summary) Webアプリケーションの開発において，開発者が頭の中で描いたイメージに従ってソフトウェアを開発することを可能とする，グラフィックスを用いたイメージ指向プログラミング技法について述べる． (Link to Search Site for Scientific Articles) ● CiNii @ National Institute of Informatics (CRID): 1572543026615871232 (CiNii: 1572543026615871232)

Review, Commentary:

1.	Kenji Ikeda : Non-Aymptotic Confidence Regions in System Identification, Systems, Control and Information, Vol.59, No.5, 192-197, May 2015. (Keyword) system identification / Confidence Regions / subsampling
2.	Kenji Ikeda : Variance Analysis of Subspace Identification Methods, Journal of the Society of Instrument and Control Engineers, Vol.50, No.11, 956-962, Nov. 2011. (Keyword) subspace identification method / bias / variance (Link to Search Site for Scientific Articles) ● CiNii @ National Institute of Informatics (CRID): 1390290316512958976 (CiNii: 1390290316512958976)
3.	Kenji Ikeda, Masami Mohri, Yoshio Mogami, Hitoshi Tokushige, Akinori Tsuji and Hiroaki Ogata : Creativity development at the University of Tokushima, Journal of the Society of Instrument and Control Engineers, Vol.46, No.1, 31-36, Jan. 2007. (Keyword) creativity development / software design / embedded system design
4.	Takao Shimomura, 高橋宗雄, Kenji Ikeda and Yoshio Mogami : Webプログラミングフレームワーク, Journal of Information Processing Society of Japan, Vol.45, No.12, 1266-1271, Dec. 2004. (Summary) 通常のプログラムとは異なり，一般に，Webアプリケーションでは，Webブラウザからのフォームデータの受け取り，値の検証，エラーメッセージの表示，セッションの管理，データベースの操作，ビジネスロジック，ページの遷移，Webブラウザに返すページの生成等の各種の処理が必要となる．これらの処理の記述を容易にし，かつ，サーバソフトウェアの保守を容易にするWebプログラミング·フレームワークについて解説し，Webプログラミング手順とそれをサポートする開発·実行環境について述べる．
5.	Kenji Ikeda : Adaptive control using logic-based switching, Journal of the Society of Instrument and Control Engineers, Vol.38, No.3, 182-188, Mar. 1999. (Summary) 従来のパラメータ適応制御の問題点である, 事前情報に関する制約を緩和すべく, 切り替え型の適応制御が提案されている. この分野の第一人者である A. S. Morse教授の解説記事に基づき, pre-routed switching, hysteresis switching, dwell-time switching などの手法の説明や安定性の原理について解説した.
6.	Yoshitomo Hanakuma, Naoki Mizuno, Kenji Ikeda and Kazuya Ogata : The future of control theory: viewed from applications, Journal of the Society of Instrument and Control Engineers, Vol.38, No.1, 65-70, Jan. 1999. (Summary) 制御理論を実用化するには, 理論面のみならず信頼性, 安全性, 経済性など難しい問題を解決せねばならない. そこで, 計測自動制御学会「適応制御部会」では,「実例を通して制御理論の今後を探る」という講義会を2回開催した. 本稿は, 企画やパネラーとして中心的な役割を果たした著者たちが, 理論と実際のギャップを明確にし, 適応制御理論の抱える問題点や, 実システムへ適用する際の問題点などを整理し, 適応制御理論研究の今後の方向性を占う内容となっている.

Proceeding of International Conference:

1.	Michiru Katayama, Kenji Ikeda and Tetsushi Ueta : Laminar stabilization control in chaos intermittency, Proc. NOLTA2023, 17-20, Catania, Sep. 2023. (Summary) Intermittency gives the irregular switching of almost periodic (laminar) and chaotic (burst) motions, and can be confirmed in every physical system within a certain parameter range. In this paper, by utilizing a method stabilizing a periodic motion which have been vanished by a saddle node bifurcation, a laminar motion in the intermittency is stabilized by a tiny control input. Apseudo periodic orbit composed by the vanished orbit or laminar motion is applied as a reference for the linear feedback control. We demonstrate the control for a circuit model and show robust stabilization responses. (Link to Publication Site) ● Publication site (DOI): 10.34385/proc.76.A2L-15 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.34385/proc.76.A2L-15 (DOI: 10.34385/proc.76.A2L-15)
2.	Kenji Ikeda and Hideyuki Tanaka : Frequency Domain Uncertainty Estimation in a Subspace Identification Method by using Covariance and Pseudo-Covariance, Proceedings of SICE 2023, 720-723, Tsu, Sep. 2023. (Summary) This paper proposes a method for estimating the uncertainty in the frequency domain of the estimate obtained by the authors' previously proposed identification method. The estimation of the pseudo-covariance matrix as well as the covariance matrix of the transfer function matrix can reduce the conservatism of the estimated uncertainty in the frequency domain. In the numerical example, the estimated standard deviations of gain and phase plots in the proposed method are compared with the standard deviations obtained from the covariance alone and the sample standard deviations of gain and phase plots and the effectiveness of the proposed method is demonstrated. (Keyword) system identification / covariance analysis / frequency domain uncertainty / Pseudo-Covariance
3.	Hideyuki Tanaka and Kenji Ikeda : State Estimation for Closed-Loop LPV System Identification via Kernel Methods, Preprints of the 22nd World Congress of the International Federation of Automatic Control, 12523-12528, Yokohama, Jul. 2023. (Summary) This paper proposes a state-estimation method for closed-loop identification of linear-parameter-varying (LPV) systems by extending an approach for linear-time-invariant (LTI) systems and applying kernel canonical correlation analysis (KCCA). The proposed method estimates the state term from the one-step-ahead prediction via a kernel approach. The incomplete Cholesky decomposition (ICD) is introduced to reduce the complexity of the KCCA. A simple numerical simulation shows the effectiveness of the proposed approach. (Keyword) LPV system identification / closed-loop identification / KCCA / Subspace methods (Link to Publication Site) ● Publication site (DOI): 10.1016/j.ifacol.2023.10.516 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85184962289 (DOI: 10.1016/j.ifacol.2023.10.516, Elsevier: Scopus)
4.	Hiroshi Oku and Kenji Ikeda : Error covariance of a closed-loop subspace model identification method for the case of general LTI beedback, Preprints of the 22nd World Congress of the International Federation of Automatic Control, 4391-4394, Yokohama, Jul. 2023. (Summary) This paper considers error-covariance analysis of a closed-loop subspace model identification method for systems in closed-loop compensated by general linear time-invariant (LTI) feedback controllers. Systems to be identified can be described with the output-error state-space representation. It is allowed that they are contaminated with an arbitrary colored output-error noise. Consistency property of our method with the presence of a colored output-error noise is also studied here. For details, since the procedure of the identification method includes the QR factorization of stacked data Hankel matrices, this study investigates asymptotic properties of block entries of the triangular matrix obtained from the QR factorization. The set of the block entries is separated into two components, namely, the signal-based component and the noise-based component. The contributions are to obtain the asymptotic covariance matrix of the vectorization of the noise-based component, and to demonstrate the consistency of the signal-based component with the presence of a colored output-error noise by a numerical simulation. (Keyword) system identification / closed-loop identification / subspace identification / error covariance analysis / consistency
5.	Kenji Ikeda and HIdeyuki Tanaka : Numerical Analysis of the Estimate of Noise Covariance by using Best Linear Unbiased Estimate, 2022 IEEE Conference on Control Technology and Applications, 109-116, Trieste, Aug. 2022. (Summary) This paper derives a BLUE (best linear unbiased estimate) for the estimation of noise covariance matrix in the framework of subspace identification. An estimation method of the noise covariance with a CCA (canonical correlation analysis) weighting matrix was previously proposed by the authors. The newly derived BLUE is compared with the previously proposed method numerically and it is shown that the previously proposed method gives a good performance. (Keyword) System identification / subspace method / best linear unbiased estimate (Link to Publication Site) ● Publication site (DOI): 10.1109/CCTA49430.2022.9966021 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85144594809 (DOI: 10.1109/CCTA49430.2022.9966021, Elsevier: Scopus)
6.	Kenji Ikeda and Hideyuki Tanaka : An Introduction of a CCA Weighting Matrix to a Closed-Loop Subspace Identification Method, Preprints of the 19th IFAC Symposium on System Identification, 467-472, (Virtual) Padova, Jul. 2021. (Summary) This paper introduces a CCA (canonical correlation analysis) weighting matrix to an estimation method of the innovations model previously proposed by the authors. A numerical simulation illustrates that the CCA weighting reduces the covariance of the estimate and that the proposed method gives similar or better performance compared to Closed-Loop MOESP and PBSID. Especially, the design parameter called "past horizon" in the proposed method can be set small compared to Closed-Loop MOESP and PBSID. It is also analyzed how the bias of the estimate is reduced. (Keyword) System identification / Subspace methods / Kalman filters / Semi-definite programming (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85118103938 (Elsevier: Scopus)
7.	Hiroshi Oku and Kenji Ikeda : On consistency of output-error closed-loop subspace model identification for systems compensated by general LTI controllers, Preprints of the 19th IFAC Symposium on System Identification, 169-173, (Virtual) Padova, Jul. 2021. (Summary) This paper deals with consistency analysis of a closed-loop subspace model identification method for systems compensated by general linear time-invariant feedback controllers. The results presented here are a straightforward extension from the authors' recent work for systems compensated with constant gain feedback controllers. It is shown that the estimate of the extended observability matrix, which is obtained from our closed-loop subspace model identification method, is consistent up to a similarity transformation (Keyword) System identification / closed loop identification / subspace methods / consistency / asymptotic properties (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85118112201 (Elsevier: Scopus)
8.	Hideyuki Tanaka and Kenji Ikeda : Minimal realization of an unstable plant under feedback, Preprints of the 19th IFAC Symposium on System Identification, 7-12, (Virtual) Padova, Jul. 2021. (Summary) This paper formulates a problem of finding a minimal realization of an unstable plant under feedback, given an infinite string of input-output data of the plant. It is assumed that the plant and noise models have non-common state matrices and that the plant is stabilized by a strictly proper linear-time-invariant controller of the finite order. A minimal realization of the innovation form has been estimated using closed-loop subspace identification algorithms, whereas that of the plant is determined in this paper. A matrix is proposed to avoid a numerical problem of a traditional algorithm applied to an unstable plant. (Keyword) closed-loop identification / joint input-output approach / unstable plant / minimal realization / stochastic realization (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85118136466 (Elsevier: Scopus)
9.	Hiroshi Oku and Kenji Ikeda : On error analysis of a closed-loop subspace model identification method, 24th International Symposium on Mathematical Theory of Networks and Systems, 701-706, Cambridge, Aug. 2020. (Summary) This paper studies error analysis and asymptotic variance of a closed-loop subspace model identification method for a system described with the output-error state-space representation. For details, since the procedure of the identification method includes the QR factorization of stacked data Hankel matrices, this study investigates asymptotic properties of block elements of the triangular matrix obtained from the QR factorization. The set of the block elements is separated into two components, namely, the signal-based component and the noise-based component. The contributions are to derive asymptotic properties of both components and to obtain the asymptotic covariance matrix of the vectorization of the noise-based component. (Keyword) System identification / subspace method / closed-loop identification / asymptotic properties / covariance matrix (Link to Publication Site) ● Publication site (DOI): 10.1016/j.ifacol.2021.06.132 (Link to Search Site for Scientific Articles) ● Search Scopus @ Elsevier (DOI): 10.1016/j.ifacol.2021.06.132 (DOI: 10.1016/j.ifacol.2021.06.132)
10.	Hiroshi Oku and Kenji Ikeda : Consistency analysis of the extended observability matrix of output-error closed-loop subspace model identification, Preprints of the 21st IFAC World Congress, 1-4, Berlin, Jul. 2020. (Summary) This paper studies statistical properties of a closed-loop subspace model identification method for a system described with the output-error state-space representation. For details, the limit value of the matrix to be singular-value-decomposed to estimate the extended observability matrix is investigated. The contribution is to ensure that the estimate of the extended observability matrix is consistent up to a similarity transform. (Keyword) system identification / closed loop identification / subspace identification method / consistency / asymtotic properties
11.	Hideyuki Tanaka and Kenji Ikeda : Subspace identification algorithm for stochastic systems equipped with zeros close to unit circle, Preprints of the 21st IFAC World Congress, 2463-2468, Berlin, Jul. 2020. (Summary) In identifying a stochastic system possessing zeros close to the unit circle, the effect of the initial state appears in the estimates. This paper derives a stochastic subspace identification algorithm for such a system. A new stochastic realization algorithm is developed based on the covariance matrices of the state and the white-noise input, by taking the initial state and positive realness into account. A subspace identification algorithm is obtained by applying the realization algorithm to a finite string of data. Numerical simulation results show that the proposed algorithm provides favorable results compared with the conventional ones. (Keyword) stochastic realization / stochastic subspace identification / positive realness / initial state / Kalman gain (Link to Publication Site) ● Publication site (DOI): 10.1016/j.ifacol.2020.12.046 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85105064752 (DOI: 10.1016/j.ifacol.2020.12.046, Elsevier: Scopus)
12.	Kenji Ikeda and Hideyuki Tanaka : An Estimation Method of Innovations Model in Closed-Loop Environment with Lower Horizons, Preprints of the 21st IFAC World Congress, 907-912, Berlin, Jul. 2020. (Summary) This paper proposes an estimation method of the innovations model in closed loop environment by using the estimate of the innovations process. The estimate of the innovations process from the finite interval of data has a bias, so are the estimate of the proposed method. However, it is analyzed that the bias can be reduced. The Kalman gain and the covariance of the innovations process are estimated by using a semi-definite programming problem previously proposed by the authors. Numerical simulation illustrates the proposed method gives better performance than Closed-Loop MOESP and PBSID when the data length is large and the past horizon is selected low. (Keyword) system identification / subspace identification method / Kalman filter / semi-definite programming (Link to Publication Site) ● Publication site (DOI): 10.1016/j.ifacol.2020.12.848 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85107612968 (DOI: 10.1016/j.ifacol.2020.12.848, Elsevier: Scopus)
13.	Hideyuki Tanaka and Kenji Ikeda : Modeling of parameter dependent systems using the impulse response and the kernel method, Proceedings of SICE Annual Conference 2019, 532-533, Hiroshima, Sep. 2019. (Summary) This paper analyzes the covariance of the estimated system matrices in PO-MOESP method and applies the result to the calculation of the covariance of the coefficient of the transfer function obtained from the estimated system matrices. In the analysis of the covariance of the system matrices, consistent estimates of Kalman gain and the covariance of the innovations process are utilized. Unscented transformation is applied to calculate the covariance of the coefficient of the transfer function. A numerical simulation is shown to illustrate the validity of the analysis. (Keyword) subspace identification / innovations model / covariance analysis
14.	Kenji Ikeda and Hideyuki Tanaka : Covariance of the Coefficient of the Transfer Function Estimated via PO-MOESP Method, Proceedings of SICE Annual Conference 2019, 518-521, Hiroshima, Sep. 2019. (Summary) This paper analyzes the covariance of the estimated system matrices in PO-MOESP method and applies the result to the calculation of the covariance of the coefficient of the transfer function obtained from the estimated system matrices. In the analysis of the covariance of the system matrices, consistent estimates of Kalman gain and the covariance of the innovations process are utilized. Unscented transformation is applied to calculate the covariance of the coefficient of the transfer function. A numerical simulation is shown to illustrate the validity of the analysis. (Keyword) subspace identification / innovations model / covariance analysis
15.	Kenji Ikeda and Hideyuki Tanaka : On the Uniqueness of the Estimate of Innovations Model, Proceedings of the ASCC2019, 1313-1318, Kitakyushu, Jun. 2019. (Summary) In this paper, a sufficient condition for identifying the Kalman gain and the covariance matrix of the innovations process is derived. Under the condition, the innovations model can uniquely be estimated in a generic sense by solving a semidefinite programming problem proposed in the previous work of one of the authors. A numerical simulation illustrates the consistency of the proposed estimate. (Keyword) system identification / subspace identification / innovation representation / linear matrix inequality (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85069893382 (Elsevier: Scopus)
16.	Kenji Ikeda and Hideyuki Tanaka : On the Uniqueness of Spectral Density Function in an SDP Problem for the Estimation of Innovations Model, Proceedings of SICE Annual Conference 2018, 1483-1488, Nara, Sep. 2018. (Summary) The authors proposed an SDP (Semi-definite programming) formulation for the identification of the Kalman gain and the covariance matrix of the innovations process, aiming at a consistent estimation. However, the rank of a coefficient matrix of the variables does not have a column full rank, so that the uniqueness of the solution is not clear. In this paper, it shown that the Kalman gain and the covariance of the innovation can be uniquely estimated under a certain condition. (Keyword) subspace identification / consistent estimate / innovations model / linear matrtix inequality
17.	Hideyuki Tanaka and Kenji Ikeda : Closed-loop subspace identification algorithm based on finite-interval realization, Proceedings of SICE Annual Conference 2018, 1489-1492, Nara, Sep. 2018. (Summary) A closed-loop subspace identification method is developed based on the realization of finite-interval data. By making a new assumption, it is shown that the proposed method shorten the period required for the one developed earlier by the authors. Numerical simulation shows effectiveness of the proposed method. (Keyword) closed-loop identification / subspace identification / Kalman gain / finite-interval realization
18.	Hideyuki Tanaka and Kenji Ikeda : Closed-loop subspace identification taking initial state into account, 18th IFAC Sympotium on System Identification (SYSID 2018), 604-609, Stockholm, Jul. 2018. (Summary) This paper develops a closed-loop subspace identification algorithm for the systems with the stochastic part possessing zeros close to the unit circle. The proposed algorithm estimates the system under the feedback loop, taking the effect of the initial state into account. A nonlinear optimization technique is applied for computing the Kalman gain and the covariance of the innovation. Numerical simulation shows that the proposed algorithm successfully provides models by taking the initial effect. (Keyword) closed-loop subspace identification / stochastic subspace identification / stochastic realization / Kalman gain / covariance of innovation (Link to Publication Site) ● Publication site (DOI): 10.1016/j.ifacol.2018.09.229 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85054430235 (DOI: 10.1016/j.ifacol.2018.09.229, Elsevier: Scopus)
19.	Kenji Ikeda and Hideyuki Tanaka : An SDP Formulation for Consistent Estimate of Innovations Model, The 2017 Asian Control Conference, 1772-1777, Gold Coast, Dec. 2017. (Summary) This paper proposes an SDP (Semi Definite Programming) problem formulation aiming at consistently estimating Kalman gain and the covariance of the innovations process. In the previous papers, the problem of estimating these parameters was formulated as a nonlinear optimization problem. It is shown that the problem can be reformulated as an SDP by extending the variables. A numerical example illustrates the consistency of the estimates. (Keyword) Identification and Estimation / LMI (Link to Publication Site) ● Publication site (DOI): 10.1109/ASCC.2017.8287442 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85047485746 (DOI: 10.1109/ASCC.2017.8287442, Elsevier: Scopus)
20.	Kenji Ikeda : An LMI Problem for Consistent Estimate of Innovations Model, Proceedings of SICE Annual Conference 2017, 587-590, Kanazawa, Sep. 2017. (Summary) This paper proposes an LMI problem for consistently estimating Kalman gain and the covariance of the innovations process. In the previous papers, the problem of estimating Kalman gain and covariance consistently is formulated as a nonlinear optimization problem. It is shown that the problem can be reformulated as an LMI problem by extending the variables and that the solution gives consistent estimates. It is also shown a sufficient condition for the existence of a consistent estimate. (Keyword) modeling / system identification / estimation (Link to Publication Site) ● Publication site (DOI): 10.23919/SICE.2017.8105577 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85044213594 (DOI: 10.23919/SICE.2017.8105577, Elsevier: Scopus)
21.	Hideyuki Tanaka and Kenji Ikeda : Identification of Linear Stochastic Systems Taking Initial State into Account, Proceedings of SICE Annual Conference 2017, 583-586, Kanazawa, Sep. 2017. (Summary) This paper derives an algorithm for identifying linear stochastic systems, by taking the effect of the initial state into account. The proposed algorithm computes coefficients by using the extended observability matrix for deleting the effect. Numerical simulation shows that the proposed method outperforms existence methods. (Keyword) modeling / system identification / estimation (Link to Publication Site) ● Publication site (DOI): 10.23919/SICE.2017.8105553 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85044142976 (DOI: 10.23919/SICE.2017.8105553, Elsevier: Scopus)
22.	Tomoaki Kashiwao, Yukari Yoshikawa, Takeshi Otsuka, Kenji Ikeda and Alireza Bahadori : Investigation of Effects of Silica Particles on Light Extraction Efficiency of Light-Emitting Diode Packaging Using Ray-Tracing Simulation, Proceedings of SICE Annual Conference 2017, 126-130, Kanazawa, Sep. 2017. (Summary) This paper reports an investigation of the effects of silica particles contained in mold resin to the light extraction efficiency of light-emitting diode (LED) packaging (PKG) using ray-tracing simulations. The light extraction efficiencies, namely, luminous and radiant flux of the rays emitted from the PKG, are improved with the silica particles. The silica particles cause the scattering of the rays in the mold resin of the LED PKG. If the scattering can be generated properly, the light extraction efficiency can be improved. Varying the volume percentage of the silica particles, its effects to the light extraction efficiency are confirmed. Finally, we could confirm the improvement of luminous higher than 0.5 lm and radiant flux higher than 0.02 W respectively from the simulation results. (Link to Publication Site) ● Publication site (DOI): 10.23919/SICE.2017.8105751 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-85042718067 (DOI: 10.23919/SICE.2017.8105751, Elsevier: Scopus)
23.	Tomoaki Kashiwao, Mayu Hiura, Bahadori Alireza, Kenji Ikeda and Takao Shimomura : Optimization of a Light Reflector Angle of Surface-mount-devide Light-emitting Diode Packaging Using Ray-tracing Simulation, Proceedings of SICE Annual Conference 2016, 361-364, Tsukuba, Sep. 2016. (Summary) This paper presents an investigation of the optimization of the light reflector angle of top-view (TV) packaging (PKG) of a surface-mount-device (SMD) light-emitting diode (LED) using ray-tracing simulations. The light extraction efficiency which is the total relative intensity of the rays emitted from the PKG, is used for the evaluation index of the output performance of the PKG. The cavity angles composing the light reflectors of TV PKG with round and square-shaped cavity are optimized on the basis of the optical design method proposed in our previous study. In addition, the optimal angles under the conditions where the mold resin contains a phosphor affecting the light extraction efficiency intensively, are investigated. Finally, we suggest important factors for designing the light reflector (cavity) angle of TV LED PKG.
24.	Hideyuki Tanaka and Kenji Ikeda : An equation for estimating Kalman gain based on the finite-interval stochastic realization, Proceedings of SICE Annual Conference 2016, 186-189, Tsukuba, Sep. 2016. (Summary) This paper derives an equation for estimating the Kalman gain based on the finite-interval stochastic realization. It also proposes an algorithm for better estimates for the Kalman gain, by improving the finite-interval realization algorithm. Numerical simulation shows that the proposed algorithm successfully improves estimates of the Kalman gain. (Keyword) stochastic realization / Kalman gain / finite interval realization / stochastic subspace identification
25.	Kenji Ikeda : Consistent Estimates of Kalman Gain and the Covariance of the Innovations Process, Proceedings of SICE Annual Conference 2016, 182-185, Tsukuba, Sep. 2016. (Summary) A problem of estimating Kalman gain and the covariance of the innovations process consistently was proposed by the author, which is formulated as a BMI problem and an algorithm solving the problem is also proposed. However, a given numerical solution requires computational costs. In this paper, in order to reduce the computational costs, an algorithm by using a Cholesky factor of some positive definite matrix is proposed. (Keyword) Subspace Identification / Consistent Estimate / Kalman gain
26.	Kenji Ikeda : Consistent Estimate of Innovations Model, Preprints, 12th IFAC International Workshop on Adaptation and Learning in Control and Signal Processing, Vol.49, No.13, 129-134, Eindhoven, Jun. 2016. (Summary) This paper proposes a numerical solution of the consistent estimates of Kalman gain and the covariance of the innovations process in the innovations model. The estimates are given as solutions of some BMI problem, so an iterative algorithm for solving the problem will be required. An algorithm for a class of BMI problems in which the objective function is nonincreasing is applied for solving this problem. Explicit formulations of the estimation error in PO-MOESP method are also derived and shown to be consistent so that the consistent estimate of all the parameters in the innovations model are obtained. (Keyword) system identification / Subspace methods / consistency / Kalman filters / nonlinear programming / Numerical solutions (Link to Publication Site) ● Publication site (DOI): 10.1016/j.ifacol.2016.07.939 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-84991043292 (DOI: 10.1016/j.ifacol.2016.07.939, Elsevier: Scopus)
27.	Kenji Ikeda : Frequency Domain Uncertainty in PO-MOESP Method, Proceedings of the 47th ISCIE International Symposium on Stochastic Systems Theory and Its Applications, 104-111, Honolulu, Dec. 2015. (Summary) This paper provides an estimation method of the uncertainty of the pulse transfer function from the covariance matrices of the estimates of the vectorized system matrices in PO-MOESP method, in which Kalman gain and the covariance of the innovations process are consistently estimated by using the previously proposed method by the author. In the conventional methods, the estimates of Kalman gain and the covariance of the innovations process are asymptotically consistent, i.e., the estimates converges in probability to the true value as not only the data length but also the past horizon go to infinity. The contribution of the paper is to provide a consistent estimate of the frequency domain uncertainty for fixed past/future horizons.
28.	Kenji Ikeda : Consistent Estimate of Kalman Gain in Subspace Identification Method, The 2015 IEEE Control Systems Society; Multiconference on Systems and Control, 151-156, Sydney, Sep. 2015. (Summary) This paper proposes a consistent estimates of the Kalman gain and the covariance of the innovations in the framework of subspace identification. The estimates of the system matrices in PO-MOESP method are known to be consistent while the estimates of (A,C) matrices in N4SID method are asymptotically consistent, i.e., each estimate converges in probability to the true value if not only the data length but also the past horizon goes to infinity. The conventional estimates of the Kalman gain and the innovations covariance are asymptotically consistent. To consider the gap relating amount, consistent estimates of the Kalman gain and the innovations covariance are obtained. (Keyword) Subspace Identification / Kalman gain / consistent estimate (Link to Publication Site) ● Publication site (DOI): 10.1109/CCA.2015.7320625 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-84964318398 (DOI: 10.1109/CCA.2015.7320625, Elsevier: Scopus)
29.	Kenji Ikeda and Hiroshi Oku : Estimation Error Analysis of System Matrices in Some Subspace Identification Methods, Proceedings of 10th Asian Control Conference, 1-6, Kota Kinabalu, Jun. 2015. (Summary) Explicit formulation of dominant parts of the estimation errors of system matrices in PO-MOESP method and a robust N4SID like method are derived based on a proposing lemma on a gap between two singular subspaces. By considering the gap instead of the perturbations of the singular vectors, it is shown that the estimate in PO-MOESP method is unbiased while the estimate in the robust N4SID like method has an asymptotic bias when the past horizon p is finite. (Keyword) system identification / analysis of variance (Link to Publication Site) ● Publication site (DOI): 10.1109/ASCC.2015.7244557 (Link to Search Site for Scientific Articles) ● Summary page in Scopus @ Elsevier: 2-s2.0-84957629036 (DOI: 10.1109/ASCC.2015.7244557, Elsevier: Scopus)
30.	Tomoaki Kashiwao, Mayu Hiura, Daigo Hara, Kenji Ikeda and Takao Shimomura : Effect of Optical Properties of Components for Light Extraction Efficiency of SMD LEDs, SICE Annual Conference 2014, 1936-1940, Sapporo, Sep. 2014. (Summary) This paper reports effect of optical properties of components for light extraction efficiency of surface mount device (SMD) Light emitting diodes (LEDs). Improvement of light extraction efficiency of SMD LEDs is very important because SMD LEDs are used for various electronic instruments. In particular, side view (SV) LEDs are used for a back light of a liquid crystal display (LCD) widely. A cavity of SV LED packaging (PKG) is narrower and frequency of reflection of light is higher than those of general square SMD LED PKG. Therefore, light extraction efficiency of a SV LED is affected by optical properties of components intensively. In the paper, effect of reflection methods of a PKG resin and silver-plating of a lead frame (L/F) is investigated with using ray tracing simulation. In addition, simulation results of SV LEDs and of square SMD LEDs are compared.
31.	Kenji Ikeda : Error Analysis of State Approaches in Subspace Identification Methods, SICE Annual Conference 2014, 1685-1690, Sapporo, Sep. 2014. (Summary) Explicit formulation of dominant parts of the estimation errors of A and C matrices in shift invariance approach and state approach are derived based on a proposing lemma on perturbations to the singular subspaces. In the state approach, two methods are considered: one estimates the state from the left singular subspaces of a certain matrix, while the other estimates the state from the right singular subspaces of the same matrix. It is shown that the estimate in the state approach has an asymptotic bias when the past horizon p is finite, while the estimate in the shift invariance approach is unbiased.
32.	Kenji Ikeda : On the precision of the plant estimates in some subspace identification methods, Preprints of the 19th World Congress IFAC, 9516-9521, Cape Town, Aug. 2014. (Summary) Explicit formulae of dominant parts of the estimation errors of plant A matrix in PI-MOESP, PO-MOESP, and one of N4SID methods are derived based on a proposing lemma on perturbations to the singular subspaces. By considereing a gap between the true and a perturbed singular subspaces instead of the perturbation on the subspace, a simple formulae of the estimation error is obtained. The order of the subordinate terms are also analysed.
33.	Kenji Ikeda : Error Analysis of PO-MOESP Algorithm, 21th International Symposium on Mathematical Theory of Networks and Systems (MTNS), 545-550, Groningen, Jul. 2014. (Summary) Explicit formulae of dominant parts of the estimation errors of A and C matrices in PO-MOESP are derived based on a proposing lemma on perturbations to the singular subspaces. The formulae are derived directly from a slightly modified PO-MOESP algorithm. Based on the formulae, the effect of the weighting matrices in the singular value decomposition is analyzed. It is also shown that the original PO-MOESP method is optimal in a sense that singular values of some matrix in the Frobenius norm of the estimation error are minimized. (Keyword) system identification / analysis of variance / Subspace model identification
34.	Takao Shimomura and Kenji Ikeda : Waiting Blocked-Tree Type Deadlock Detection, Proc. of Science and Information Conference 2013 (SAI2013), 45-50, London, Oct. 2013.
35.	Tomoaki Kashiwao, Kenji Ikeda and Takao Shimomura : Identification of a Hopping Machine Based on a Piecewise Affine Model, SICE Annual Conference 2013, 501-504, Nagoya, Sep. 2013. (Summary) In this study, a hopping machine (HM) is identified as a piecewise affine (PWA) system via estimating a transition time. The HM can be modeled as the PWA model with two types of discrete transitions: one caused by the change between discrete free-fall and spring-mass states and the other caused by inversion of the sign of a nonlinear drag coefficient term owing to air resistance. First, the equation of motion for the HM is obtained considering the nonlinear term in order to analyze those discrete transitions. In addition, the transition times are estimated using a wavelet-based method proposed in our previous papers. Finally, the HM is identified as a piecewise affine auto regressive exogenous (PWARX) model on the basis of the analyzed discrete states. (Keyword) piecewise affine system / transition time / identification / wavelet analysis / hopping machine
36.	Kenji Ikeda : Asymptotic Properties of MOESP-type Methods, Preprints of ALCOSP and PSYCO 2013, 92-97, Caen, Jul. 2013. (Summary) Precision of the estimated plant model is often the main interest of system identification. In order to take the predictable part of the noise into account, the noise model is estimated together with the plant model by using ARMAX model. In that case, a model reduction procedure will be required in order to obtain the plant model. On the other hand, the plant model can be estimated directly by using output error (OE) model. In this paper, PI-MOESP method and PO-MOESP method are compared by analysing the signal and noise components of the estimated plant model under the assumption that there are no common poles in the plant and the noise models. The magnitude of the noise component in each method is discussed when the past or future horizon varies and it is shown that there is a possibility that PI-MOESP method gives better performance than PO-MOESP method. (Keyword) System Identification / Subspace Model Identification / Variance Analysis
37.	Takao Shimomura, Kenji Ikeda and Montanaro Adriano : AUTOMATIC COURSEWARE TRANSFERS BETWEEN DIFFERENT PLATFORMS, Proc. of 6th International Technology, Education and Development Conference (INTED2013), 4334-4341, Valencia, SPAIN, Mar. 2013.
38.	Takao Shimomura and Kenji Ikeda : Extensible Web-based Learning Architecture, Proc. of the 2012 International Conference on Software Engineering Research and Practice (SERP'12), 1-7, Las Vegas, Jul. 2012.
39.	Kenji Ikeda and Oku Hiroshi : Asymptotic Properties of BCSS Method in Closed Loop Environment, Preprints of the 16th IFAC Symposium on System Identification, 858-863, Brussels, Jul. 2012. (Summary) Bias compensated state space model identification method in closed loop environment (CL-BCSS) is introduced. The noise is assumed to be a 0 mean colored noise. It is also assumed that the reference input is a white Gaussian process uncorrelated to the noise. Signal and noise components of a certain matrix in the proposed method and CL-MOESP method are analyzed and compared. It is shown that CL-BCSS method corresponds to CL-MOESP method with infinite past horizon. (Keyword) closed-loop identification / State-space methods / Variance
40.	Takao Shimomura, Kenji Ikeda and Montanaro Adriano : GAME-ORIENTED STIMULATING E-LEARNING, Proc. of 6th International Technology, Education and Development Conference (INTED2012), 4221-4228, Valencia, Spain, Mar. 2012.
41.	Takao Shimomura and Kenji Ikeda : Tag Command-based Automatic Trouble Diagnosis for Computer Administration, the 5-th International Conference on New Trends in Information Science and Service Science (NISS2011), 97-102, Macao, Oct. 2011.
42.	Kenji Ikeda, Yoko Hatanaka and Takao Shimomura : Linear Approximation Model for Snake Robots and its Application, Proc. of SICE Annual Conference 2011, 2915-2918, Tokyo, Sep. 2011. (Summary) This paper proposes a modeling method of snake robots and analyzes the obtained model. Regarding the snake robot as a periodic system, a discrete-time linearized approximation model is obtained based on lifting technique and Fourier series expansion. PI-MOESP type subspace model identification method is applied for obtaining the model. Hankel singular values of the obtained model are analyzed and some remarks are given for the selection of the control input. (Keyword) Periodic systems / Subspace model identification method / Lifting / snake robot
43.	Tomoaki Kashiwao, Kenji Ikeda and Takao Shimomura : A Design of the Scale Parameter of Wavelet Transformation for the Estimation of Transition Time in PWL Systems, Proc. of SICE Annual Conference 2011, 973-978, Tokyo, Sep. 2011. (Summary) This paper proposes a method to determine an appropriate scale parameter of wavelet transform for the estimation of transition time of the Piecewise Linear (PWL) systems by using wavelet analysis. Precision of the estimation of transition time depends on the scale parameter of wavelet transform under noisy environment. Effect of the noise on the estimation of transition time is shown under the assumption of Gaussian random noise in a numerical example. Appropriate range of the scale parameterof wavelet transform is determined by using quantified effect of the noise and effectiveness of the method is illustrated by numerical experiments. (Keyword) piecewise linear systems / transition time / wavelet transform / scale parameter
44.	Kenji Ikeda and Hiroshi Oku : Asymptotic Properties of Bias Compensated State Space Model Identification Method, Preprints of the 18th IFAC World Congress, 6505-6510, Milan, Italy, Aug. 2011. (Summary) The authors proposes a bias-compensated state space model identification (BCSS) method in order to relax the order of the persistent excitation (PE) condition. Most of the subspace identification methods utilize instrumental variable (IV) to obtain an unbiased estimate while the proposed method introduces a bias compensation technique to the Ordinary MOESP type method. In this paper, S/N ratios of certain matrices in the proposed method and the PI-MOESP method are analyzed and compared. It is shown that the S/N ratio in the proposed method is better than that in the PI-MOESP method especially when a pole of the plant is near 1 and the number of the block rows of the IV matrix is small. (Keyword) Identification algorithm / Subspace methods / Asymptotic analysis / S/N ratio
45.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : BCLS Method for the Estimation of Continuous-Time Models in Closed Loop Environment, Proceedings of 2010 IEEE International Conference on Control Applications, 1369-1373, Yokohama, Sep. 2010. (Summary) This paper proposes a bias-compensated least squares (BCLS) method for a continuous-time model estimation in closed loop environment. The feedback controller is assumed to be linear time invariant discrete-time system and the feedback system is assumed to be asymptotically stable. It is also assumed that the control input is a zeroth order hold signal. The aim of the proposed method is to reduce the computational cost which will be required for the conventional method. (Keyword) Parameter Estimation / Continuous-Time Model / Least Squares Method / Closed Loop Systems / Bias Compensation
46.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias Compensated State Space Identification Method in Closed Loop Environment, Proceedings of IFAC ALCOSP 2010, Antalya, Turkey, Aug. 2010. (Summary) A method of bias compensation in subspace identification method in closed loop environment is proposed. The noise is assumed to be a 0 mean colored noise and is assumed to be uncorrelated with the reference input. The covariance matrix of the noise is estimated directly from the residuals instead of estimating the noise model. The proposed method becomes an iterative algorithm and it is analyzed that the proposing method achieves 2nd order convergence. (Keyword) Closed-loop identification / State-space methods / Convergence analysis / Iterative methods / Bias compensation
47.	Takao Shimomura, Kenji Ikeda and Takahashi Muneo : An Approach to GA-driven Automatic Refactoring based on Design Patterns, 2010 Fifth International Conference on Software Engineering Advances, 213-218, Nice, Aug. 2010.
48.	Min Zheng, Kenji Ikeda and Takao Shimomura : Estimation of Continuous-time Nonlinear Systems by using the Unscented Kalman Filter, Proc. of SICE Annual Conference 2010, 2192-2195, Taipei, Aug. 2010. (Summary) This paper proposes a continuous-time model estimation method by using the Unscented Kalman Filter (UKF) from the sampled I/O data, in which the plant parameters as well as the initial state are estimated. The initial state is estimated based on a backward system of the plant, and the parameters are estimated by using the iterated UKF, which repeats the estimation of the forward system and the backward system alternately. In order to demonstrate the effectiveness of the proposed method, the rotary pendulum is provided to estimate the parameters of the continuous-time nonlinear system. (Keyword) Nonlinear Systems / System Identification / Estimation
49.	Kenji Ikeda, Tomoaki Kashiwao and Takao Shimomura : Estimation of a Transition Time in Piecewise Linear Systems, Proc. of SICE Annual Conference 2010, 620-624, Taipei, Aug. 2010. (Summary) This paper proposes an estimation method of the transition time of piecewise linear (PWL) systems by using wavelet analysis. A definition of detectability of transitions is proposed in advance and its necessary and sufficient condition is presented which gives an insight how to estimate an transition time. The proposing method does not utilize the parameters of the plant model but utilizes a characteristic pattern of the wavelet transformation of the output signal around the transition time. (Keyword) Hybrid systems / PWL systems / wavelet analysis / transition time
50.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-Compensated State Space Model Identification, Proceedings of 19th International Symposium on Mathematical Theory of Networks and Systems, 1369-1372, Budapest, Jul. 2010. (Summary) A method of bias compensation in subspace identification method is proposed. The noise is assumed to be colored with 0 mean and is assumed to be uncorrelated with the input. The covariance matrix of the noise is estimated directly from the residuals instead of estimating the noise model. The proposing method becomes an iterative algorithm but it converges with order 2. (Keyword) State Space Model Identification / Bias Compensation
51.	Takao Shimomura and Kenji Ikeda : SERVER-DRIVEN REGRESSION TEST FOR WEB APPLICATIONS AND ITS PERFORMANCE, Proceedings of the IASTED International Conference on Software Engineering, 234-239, Innsbruck, Austria, Feb. 2010.
52.	Takao Shimomura, Kenji Ikeda and Takahashi Muneo : ATTACHABLE SOFTWARE VISUALIZATION FOR MULTIPLE-THREAD PROGRAMS, Proceedings of the IASTED International Conference on Software Engineering, 135-141, Innsbruck, Austria, Feb. 2010.
53.	Takao Shimomura, Kenji Ikeda and Muneo Takahashi : Synchronization of Multi-window requests for Server-side Regression Test of Web Applications, Proceedings of the 9th International Conference on Quality Software (QSIC2009), 1-6, Jeju, Korea, Aug. 2009.
54.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-Compensated Least Squares Method in Closed Loop Environment, --- Estimation under colored noise ---, Proceedings of ICCAS-SICE 2009, 4923-4926, Fukuoka, Aug. 2009. (Summary) In this paper, bias-compensated least squares (BCLS) method is proposed in order to estimate a plant under feedback control when the noise is colored. The feedback controller is assumed to be linear time invariant discrete-time system and the feedback system is assumed to be asymptotically stable. The aim of the proposed method is to reduce the computational cost which will be required for the conventional method. (Keyword) bias compensation / least squares method / colored noise
55.	Takao Shimomura, Chen Liang Quan and Kenji Ikeda : Customizable Functional Web Components for Visual Web Programming, 2009 WRI World Congress on Computer Science and Information Engineering, 580-584, Los Angeles, Apr. 2009.
56.	Takao Shimomura, Kenji Ikeda and Takahashi Muneo : Web-Based Troubleshooter Generation for System Administration, 2009 WRI World Congress on Computer Science and Information Engineering, 760-764, Los Angeles, Apr. 2009.
57.	Takao Shimomura, Chen Liang Quan and Kenji Ikeda : SERVER-SIDE REGRESSION TEST FOR WEB APPLICATION DEVELOPMENTS, Proceedings of the IASTED International Conference on Software Engineering, 63-68, Innsbruck, Feb. 2009.
58.	Takao Shimomura, Kenji Ikeda and Muneo Takahashi : SOFTWARE VISUALIZATION WITH ATTACHABLE DISPLAY CLASSES, Proceedings of the IASTED International Conference on Software Engineering, 37-42, Innsbruck, Feb. 2009.
59.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias Compensation of Recursive Least Squares Estimate in Closed Loop Environment, Proceedings of the 18th International Symposium on Mathematical Theory and Networks and Systems, Virginia, USA, Jul. 2008. (Summary) In this paper, an asymptotic bias of the recursive least squares (RLS) estimate in the closed loop enviornment is analyzed and its compensation method is proposed under the assumption that the noise is white. Namely, a bias compensated RLS method in the closed loop environment based on output error (OE) model is proposed. A posteriori error is also analyzed for the estimation of the noise variance. (Keyword) Recursive Least Squares Estimate / Bias Compensation / Closed Loop Identification
60.	Kenji Ikeda : Bias-compensated Least Squares Method in Closed Loop Environment, Proc. of IFAC WC 2008, 13379-13384, Seoul, Jul. 2008. (Summary) In this paper, a bias-compensated least squares(BCLS) method in the closed loop environment is proposed. It is assumed that the observation noise is a white gaussian signal while there are no process noises. It is also assumed that the plant is controlled by a linear time invariant controller and that the closed loop system is asymptotically stable. The proposed estimator is unbiased and it does not require the reference input to be informative. An iterative redesign of the prefilter is also considered to achieve a minimum variance estimator. The proposed BCLS method is applied for the iterative redesign of the prefilter in order to reduce the computational cost. (Keyword) bias compensation / least squuares estimator / closed loop identification
61.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-Compensated Least Squares Method in Closed Loop Environment, --- A filter design for colored noise ---, Proceedings of 2008 International Symposium on Advanced Control of Industrial Processes, 329-334, Jasper, May 2008. (Summary) In this paper, a bias-compensated least squares (BCLS) method in the closed loop environment is extended for the case where the measurement noise is colored. Namely, a BCLS method in the closed loop environment based on Box-Jenkins model is proposed. Noise model is estimated based on a non-parametric model or high order model because the purpose of the closed loop identification is the improvement of the plant dynamics and the purse of the identification of the noise model is to obtain the unbiased plant parameters with smaller variances. (Keyword) Bias compensated least squares method / closed loop identification
62.	Takao Shimomura, Lang Sok Nhor and Kenji Ikeda : Extensible Syntax-Oriented Verifier and Its Implementation, Proceedings of the IASTED International Conference on Software Engineering, 298-303, Innsbruck, Austria, Feb. 2008.
63.	Kashiwao Tomoaki, Kenji Ikeda and Takao Shimomura : Estimation of the transition time of hybrid systems by using wavelet analysis, Proc of SICE Annual Conference, 2546-2549, Takamatsu, Sep. 2007. (Summary) Estimation of transition time is important for identification of hybrid systems. This paper proposes an estimation method of the transition time of the hybrid system by using wavelet analysis. Discrete transition of the state can be detected as a characteristic pattern in high frequency range. The effectiveness of the proposed method is illustrated by numerical experiments. (Keyword) hybrid systems / wavelet analysis / estimation
64.	Zheng Min, Kenji Ikeda and Takao Shimomura : Estimation of Rotary Inverted Pendulum by using the Unscented Kalman Filter, --- Estimation of the initial state ---, Proceedings of SICE Annual Conference, 1670-1673, Takamatsu, Sep. 2007. (Summary) This paper proposes a continuous-time model estimation method by using the Unscented Kalman Filter (UKF) from the sampled I/O data, in which the initial state as well as the plant parameter are estimated. The initial state is estimated based on a backward system of the plant. The effectiveness is demonstrated by using the Rotary Inverted Pendulum (RIP) to estimate the initial state and the parameter of a continuous-time nonlinear system. (Keyword) nonlinear system / Unscented Kalman Filter / continuous-time model
65.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-compensated adaptive observer for a continuous-time model estimation, Preprints 9th IFAC Workshop Adaptation and Learning in Control and Signal Processing, St. Petersburg, Aug. 2007. (Summary) In this paper, a bias-compensating method for a continuous-time model estimation by using adaptive observer is proposed. It is assumed that the observation noise is a white gaussian signal while there are no process noises. The proposed method is applicable for the identification in the closed loop environment. (Keyword) continuous-time systems / parameter estimation / closed loop systems / bias compensation / adaptive filters
66.	Takao Shimomura, Chen Liang Quan, Lang Sok Nhor and Kenji Ikeda : Integrated Laboratory Network Management System, Proceedings of the 7th WSEAS International Conference on Applied Informatics and Communications, 188-193, Athens, Aug. 2007.
67.	Takao Shimomura, Kenji Ikeda, Chen Liang Quan, Lang Sok Nhor and Takahashi Muneo : Apty: Easy Enjoyable Effective e-Learning, Proceedings of the 7th WSEAS International Conference on Applied Informatics and Communications, 211-216, Athens, Aug. 2007.
68.	Kenji Ikeda : Bias-compensating method for a continuous-time model estimation by using adaptive observer, Proceedings of European Control Conference 2007, 1517-1522, Kos, Jul. 2007. (Summary) In this paper, a bias-compensating method for a continuous-time model estimation by using adaptive observer is proposed. It is assumed that the observation noise is a white gaussinan signal while there are no process noises. The proposed method is applicable for the estimation of unstable plants in closed loop environment. (Keyword) continuous-time systems / parameter estimation / closed loop systems / bias compensation
69.	Takao Shimomura, Kenji Ikeda, Chen Ling Quan, Lang Sok Nhor and Takahashi Muneo : Visual Pivot-Table Components for Web Application Development, Proc. of 3rd IASTED International Conference on Advances in Computer Science and Technology, 90-95, Phuket, Apr. 2007.
70.	Takao Shimomura, Kenji Ikeda, Chen Liang Quan, Lang Sok Nhor and Takahashi Muneo : Customizable Verifiers for Web Applications and Their Implementation, Proceedings of International Conference on Computer Engineering and Applications, 396-401, Gold Coast, Australia, Jan. 2007.
71.	Takao Shimomura, Kenji Ikeda, Chen Liang Quan, Lang Sok Nhor and Takahashi Muneo : Rich Component Generation for Web Applications Using Custom Tags, Proceedings of International Conference on Computer Engineering and Applications, 390-395, Gold Coast, Australia, Jan. 2007.
72.	Takao Shimomura, Kenji Ikeda, Chen Liang Quan, Lang Sok Nhor and Takahashi Muneo : Visual Programming for Web Applications that Use HTML Frame Facilities, Proceedings of International Conference on Computer Engineering and Applications, 384-389, Gold Coast, Australia, Jan. 2007.
73.	Min Zheng, Kenji Ikeda and Takao Shimomura : Parameter estimation of rotary inverted pendulum based on unscented Kalman filter, Proc. of SICE-ICCAS 2006, 5022-5025, Busan, Oct. 2006. (Summary) The Unscented Kalman Filter (UKF) is a nonlinear estimation method, and has been widely applied in nonlinear control applications. However, if covariance of the initial value is unknown, the estimated value does not converge to the true value in general. To improve the result of estimation, the iterated UKF is propsed as a method to repeat the estimation by redefining the initial value of the filter. Its effectiveness is demonstrated by using the Rotary Inverted Pendulum(RIP) to estimate the parameters of a continuous-time nonlinear system. (Keyword) continuous-time nonlinear system / Unscented Kalman Filter / Rotary Inverted Pendulumn / estimationi
74.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-time model identification by using adaptive observer, --- Estimation of the initial state ---, Proc. of SICE-ICCAS 2006, 1796-1799, Busan, Oct. 2006. (Summary) Importance of the continuous-time model identification is increasing as the computer becomes faster. The main difference between the contiuous-time model identification and the discrete-time identification lies not only in the information loss incurred by sampling but also in that the transient term must be taken into account in the continous-time identification. This paper proposes a continous-time model estimation method from sampled I/O data with the ZOH input assumption. The initial state of the plant is estimated by using a backward system of the plant. An upper bound of the plant degree is assumed to be known. Numerical example shows that the proposed method performs well as the conventional method with an initial state estimation. (Keyword) continous-time systems / parameter estimation / sampled-data systems / adaptive filters
75.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Parameter convergence of a continuous-time model estimation method by using adaptive observer, Proceedings of 17th International Symposium onMathematical Theory of Networks and Systems, 1096-1101, Kyoto, Jul. 2006. (Summary) In this paper, a continous-time model estimation method by using adaptive observer from sampled I/O data with the ZOH input assumption is considered. The proposed method utilizes modified state variable filters in order to estimate the intersample signals. A sufficient condition for the exponential convergence of the parameter estimate is analyzed. (Keyword) continuous-time systems / parameter estimation / sampled-data systems / adaptive filters
76.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Discrete-time state variable filters for the estimation of a continuous-time model, Proceedings of the 14th IFAC Symposium on System Identification, 861-866, Newcastle, Australia, Mar. 2006. (Summary) This paper proposes a continuous-time model estimation method from sampled I/O data with the ZOH input assumption. State variable filters are modified in order to estimate the intersample signals by using a previously obtained parameter estimate. A sufficient condition for the parameter convergence and the bias of the estimated parameters are analyzed. (Keyword) continuout-time systems / parameter estimation / sampled-data systems
77.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-Time Model Identification by using Adaptive Observer, --- On the reduction of the computational cost ---, Proceedings of SICE 2005, 246-251, Okayama, Aug. 2005. (Summary) The authors have proposed a continuous-time model identification method by using adaptive observer which estimates the intersample output and the initial state of the plant as well as the plant parameters from sampled I/O data. In the previously proposed method, a numerical simulation technique such as Runget-Kutta method is used in order to estimate the intersample signals. However, the continuous-time simulation requires the high computational cost compared to the discrete-time simulation. In this paper, the discrete-time simulation technique is adopted even for the continuous-time state variable filters in order to reduce the computational cost. (Keyword) continuous-time model / adapitve observer / identification
78.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-Time Model Identification using Adaptive Observers, --- On the Selection of State Variable Filters ---, Proceedings of the 24th IASTED International Conference Modelling, Identification, and Control, 160-165, Innsbruck, Feb. 2005.
79.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : A sufficient condition for the convergence of adaptive observer in continuous-time model identification, Proceedings of IFAC Workshop on Adaptive and Learning in Control and Signal Processing, 699-704, Yokohama, Aug. 2004. (Summary) This paper proposes a new method to identify a continuous-time model from sampled I/O data by using an adaptive observer. The boundedness of the parameter estimate and the exponential convergence of the parameter estimation error to 0 under the assumption of Persistently Exciting(PE) signals are guaranteed. In order to identify the plant from a finite number of I/O data, an adaptive observer for a backward system is also proposed. In noise-free cases, the magninude of the parameter estimation error can be made arbitrarily small if the sampling period is smaller than a certain value. (Keyword) continuous-time model / identification / sampled-data system / adaptive observer
80.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-Time Model Identification by using Fast Adaptive Observer, Proceedings Sixteenth International Symposium on Mathematical Theory of Networks and Systems, Leuven, Jul. 2004. (Summary) This paper introduces a new method to identify a continuous-time model from sampled I/O data by using an adaptive observer. The adaptive algorithm is based on Newton-like method that the convergence of the parameter is considereably improved compared to the steepest descent method. In order to identify the plant from a finite number of I/O data, an adaptive observer for a backward system is also proposed. (Keyword) continuous-time model / identification / sampled-data system / adaptive observer
81.	Tagami Takanori, Kawabe Tohru and Kenji Ikeda : Multi-objective Design Scheme for Robust I-PD Controller, Proc. of the 23rd IASTED International Conference on Modelling, Identification, and Control, 128-131, Grindelwald, Feb. 2004. (Summary) 筆者の一人は, 速応性と振動抑制性を考慮して, 一般化ISEに基づく制御系設計手法を提案している. 一般化ISEは, 出力誤差の2乗積分と出力誤差の微分値の2乗積分の和である. それぞれの項は, 速応性と振動抑制性に関係している. ところが, それらの項をどのような割合で足したものが評価関数としてよいのかは, 事前にはわからない. そこで, 本論文では, それぞれの項の最小化する2目的問題として定式化し, 速応性と振動抑制性のトレードオフをとる設計手法を提案している. (Keyword) Generalized ISE / I-PD controller / Genetic Algorithm / Pareto-optimality
82.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-Time Model Identification By Using Adaptive Observer, --- A case of non-uniform sampling ---, Proc. of the 23rd IASTED International Conference Modelling, Identification, and Control, 92-97, Grindelwald, Feb. 2004. (Summary) This paper proposed a continuous-time model identification from sampled I/O data by using an adaptive observer in which the output is non-uniformly sampled. The boundedness of the parameter estimate error and the exponential convergence of the parameter estimate error to 0 under the PE assumption are guaranteed. In order to identify the plant from a finite number of I/O data, an adaptive observer of a backward system is also proposed. (Keyword) adaptive observer / identification / continuous-time model / sampled-data system
83.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-time model identification by using adaptive observer, Preprints of 13th IFAC Symposium on System Identification, 325-330, Rotterdam, Aug. 2003. (Summary) This paper proposes a continuous-time model identification from sampled I/O data by using an adaptive observer. The boundedness of the parameter and the exponential convergence of the parameter estimate error to 0 under the PE assumption are guaranteed. In order to identify the plant from a finite number of the I/O data, an adaptive observer of a backward system is also proposed. (Keyword) Continuous-time model / Identification / Sampled-data system / Adaptive observer
84.	Takao Shimomura, Muneo Takahashi, Kenji Ikeda and Yoshio Mogami : An Approach to Brain-Image Oriented Programming for Web Applications, Proc. of 7th World Multiconference on Systemics, Cybernetics and Informatics, 436-441, Orlando, Jul. 2003. (Summary) ユーザが頭に描いたイメージを基にして，プログラムを開発でき，プログラムを構成するコンポーネントの完備性や，コンポーネント間の関係の一貫性をユーザが容易に検証でき，___ 設計の途中でも，ユーザがそれまでに設計したコンポーネントの実行結果を容易に確認できるという特徴をもつイメージ指向プログラミング方式について提案し，アプリケーションを対象として本方式を実現したシステムについて述べる．
85.	Takanori Tagami, Kenji Ikeda and Tohru Kawabe : Design of robust I-PD controller based on Pareto-optimality with generalized ISE, Proceedings of the 22nd IASTED International Conference on MODELLING, IDENTIFICATION, AND CONTROL, 566-571, Innsbruck, Feb. 2003. (Summary) 先の論文では, 速応性と振動抑制性の2目的最適化問題を, 閉ループ系の極の実部の最大値の最小化, および, 原点から見た虚軸からの角度の最大値の最小化という2つの最小化問題に置き換えて, 2目的問題を解いた. 一方, 筆者の一人はロバスト性を考慮した上で, 速応性と振動抑制性の両方を考慮した制御系設計手法として一般化ISEという評価指標に基づく手法を提案している. そこで, 本論文では, 二つの手法の特徴を実際の磁気浮上系を例題にとり, 比較·検討した.
86.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Adaptive Observer of Sampled-Data Systems, Proceedings of the 22nd IASTED International Conference MODELLING, IDENTIFICATION, AND CONTROL, 102-107, Innsbruck, Feb. 2003. (Summary) 与えられたサンプル値入出力データから対象の連続時間モデルを直接同定する適応観測器を提案し, その収束性いついて解析した. サンプル点間の応答を得られたプラントモデルに基づいて推定し, また, 推定された連続時間応答からプラントモデルを推定するという繰り返し型のアルゴリズムである.
87.	Takanori Tagami and Kenji Ikeda : Design of robust I-PD controller based on multi-objective genetic algorithm with local search techniques, Proceedings of the International Symposium on Nonlinear Theory and Its Applications, 543-546, Xi'an, Oct. 2002. (Summary) 先に提案したロバストI-PD制御器設計法は, 速応性と振動抑制性を制御系の整定時間と減衰比の改善問題と捉え, さらに, 制御系の極配置問題に置き換えてから遺伝的アルゴリズムによって最適解を求めるという手法である. 最適解探索アルゴリズムにおいて, 収束性を改善するために, 局所探索手法を導入した.
88.	Kenji Ikeda : Identification of the jump in switched linear systems, Proceedings of the International Symposium on Nonlinear Theory and Its Application, 535-538, Xi'an, Oct. 2002. (Summary) ハイブリッドシステムの一種である区分線形システムの動特性を同定する場合, 連続的な動特性だけでなく, 離散的な状態遷移についても同定する必要がある. 本論文では, 離散的な状態遷移が起こる時刻は既知として, 離散的な状態遷移と連続的な動特性を同時に推定する手法を提案し, 数値例によって, 有効性を検証した.
89.	Takanori Tagami and Kenji Ikeda : A Design of Robust I-PD Controller Based on Genetic Algorithm, Proceedings of the IASTED International Conference on MODELLING, IDENTIFICATION, AND CONTROL, 290-295, Innsbruck, Feb. 2002. (Summary) A New design methos of robust I-PD controller based on Genetic Algorithm(GA) is presented. It is defined as a two-objective optimization problem in which an issue of settleing time and damping ratio is translated into a pole assignment problems. The uncertainty of the plant is represented as a polytope of polynomials and a design cost is reduced by using the edge theorem.
90.	Kenji Ikeda : Limit Cycle in Fault Tolerant Decentralized Adaptive Control System, Proceedings of 2000 International Symposium on Nonlinear Theory and its Applications, 31-34, Dresden, Sep. 2000. (Summary) 筆者らが提案している自律分散適応制御において, あるサブシステムが故障したとき, 解軌道のリミットサイクルが, 故障後に別のリミットサイクルに変形することによって, 耐故障性が達成されていることを示した.
91.	Kenji Ikeda and Seiichi Shin : Autonomous Decentralized Adaptive Control System Using Parameter Projection, Proceedings of AFSS 2000, 293-298, Tsukuba, May 2000. (Summary) 筆者らが提案している分散型ハイゲイン適応制御は, パラメータ調整則にσ修正法を用いているため, 雑音や外乱が存在しなくても, 出力が0には収束しない場合がある. そこで, パラメータがある範囲内ならばσ修正項を無効にする調整則を提案し, 安定性の解析, 耐故障性を示した. また, σ修正項が無効となる範囲を適切に選ぶことにより, 出力の0への収束性を数値例により確認した.
92.	Kenji Ikeda and Shin Seiichi : Adaptive Stabilization of Any Minimum Phase Plant with Relative Degree n or n+1, Preprints of the 11th IFAC Symposium on System Identification, Kitakyushu, Jul. 1997. (Summary) 正実関数には+1,0,-1の相対次数に関する自由度があることを利用して相対次数が正確にわかっていなくても適応的に安定化する適応制御を提案した. 相対次数が-1の伝達関数は非因果的となるため, 本論文では 0 と+1 に相当する場合について安定性を証明した.
93.	Kenji Ikeda and Shin Seiichi : Fault Tolerant Decentralized Adaptive Control Systems Using Backstepping, Proc. of the 34th IEEE Conf. on Decision and Control, 2340-2345, New Orleans, Dec. 1995. (Summary) Backstepping 法を用いた分散型ハイゲイン適応制御系をサーボ問題へ拡張した. これによって, ステップ目標値, 正弦波, ランプ目標値などへ対応できるようになり, また, 各サブシステムの相対次数がnである場合へ適用可能となった. 耐故障性やオンライン拡張性などのよい性質もそのまま成り立つ.
94.	Kenji Ikeda and Shin Seiichi : Decentralized High Gain Adaptive Control System Using Backstepping, Proc. of the 3rd IEEE Symposium on Emerging Technologies and Factory Automation, 257-264, Tokyo, Nov. 1994. (Summary) 筆者らが提案している分散型ハイゲイン適応制御は各サブシステムの相対次数が1であるという制約条件が必要であったが, この制約をbackstepping法を用いることによって緩和した.
95.	Kenji Ikeda : Simultaneous Stabilization by Periodically Switching Control, Proc. of the Asian Control Conference, 455-458, Tokyo, Aug. 1994. (Summary) 同一の線形補償器では安定化不可能な2つの対象が与えられたとき, 同じアルゴリズムに基づく制御によって安定化する, 同時安定化問題を考えた. 本論文では, 周期時変補償器を提案し, 安定性の解析を行った.
96.	Shin Seiichi and Kenji Ikeda : Autonomous Decentralized Adaptive Control, Proc. of IEEE 2nd International Workshop on Emerging Technologies and Factory Automation, 204-209, North Queensland, Sep. 1993. (Summary) 本論文では, 筆者らが提案している分散型ハイゲイン適応制御は耐故障性を持つことを自己相似性に基づいて示し, 分散型制御系の特徴である, 耐故障性, オンライン拡張性などを説明する原理を明らかにした.
97.	Kenji Ikeda, Shin Seiichi and Kitamori Toshiyuki : A Servo Design of Decentralized Adaptive Control, Proc. of International Symposium on the Mathematical Theory of Networks and Systems, 249-252, Regensburg, Jul. 1993. (Summary) 先に提案した分散型ハイゲイン適応制御では, 制御系の安定化を目的としていたが, 本論文では, 一定目標値などへの追従を制御目的としたサーボ系の構成へ拡張した. すべての信号の有界性と, 耐故障性の解析を行っている.
98.	Kenji Ikeda, Shin Seiichi and Kitamori Toshiyuki : Fault Tolerance of Decentralized Adaptive Control, Proc. of International Symposium on Autonomous Decentralized Systems, 275-281, Kawasaki, Mar. 1993. (Summary) 分散型制御系の特徴として, 耐故障性, 保守·点検の容易さ, オンライン拡張性などが挙げられている. 本論文では, 分散型ハイゲイン適応制御の安定性解析に基づき, 耐故障性を持つことを自己相似性に基づいて示した.
99.	Miyamoto Sadaaki and Kenji Ikeda : A Method of Fuzzy Control Using a Probabilistic Model, Proc. of the Korea-Japan Joint Conference on Fuzzy Systems and Engineering, 183-186, Seoul, Aug. 1992. (Summary) ファジィ制御は, ファジィ集合を使わずとも, 確率モデルを用いて構成可能であると思われていたが, これは証明されたことではなかった. 本論文では, まず, この予想が正しいことを示した上で, ファジィ制御の方が確率的手法に比べて使いやすいこと, および, 確率的手法では確率密度関数に関する完全な知識が必要である点を指摘した.

Proceeding of Domestic Conference:

1.	Kenji Ikeda and Hideyuki Tanaka : Covariance Analysis of the Estimated System Matrices in a Subspace Identification Method, 第11回計測自動制御学会制御部門マルチシンポジウム, 177-184, Mar. 2024. (Summary) This paper proposes an estimate of the covariance matrix of the system matrices estimated in a subspace identification method previouly proposed by the authors. The estimate of the covariance matrix can be derived from a set of input/output data used for the identification. It is important to provide a covariance matrix of the estimate in order to guarantee the quality of the estimate. (Keyword) Subspace identification / covariance analysis / Markov Parameters
2.	Kenji Ikeda and Hideyuki Tanaka : Uncertainty Analysis of Estimates in a Subspace Identification Method, 第66回自動制御連合講演会, 361-368, Oct. 2023. (Summary) This paper numerically analyzes the uncertainty of the estimates in the subspace identification method proposed by the authors. The authors have already derived an expression for the uncertainty of the system matrices. The covariance of the Markov parameters identified from open-loop data are in good agreement with the sample covariances. However, the uncertainty of those of the closed-loop identification has some discrepancy between the derived estimate of the covariance and the sample covariance. In this presentation, we analyze the cause of this discrepancy. (Keyword) Subspace identification / Innovations model / Noise covarialnce / covariance of the estimate
3.	Hideyuki Tanaka and Kenji Ikeda : Covariance Matrices of a Stochastic Linear-Time-Varying System, 第66回自動制御連合講演会, 361-368, Oct. 2023. (Summary) Innovation representations are often used to identify linear-parameter-varying (LPV) systems. However, the output is not always generated by the innovation representation. Motivated by the investigation of the innovation representation of the identification of LPV systems, the authors study the covariance matrix of a stochastic linear-time-varying system and obtain the innovation representation on a finite interval. (Keyword) stochastic linear-time-varying system / finite interval / innovation representation
4.	Hideyuki Tanaka and Kenji Ikeda : Numerical analysis of state-term estimation in closed-loop LPV identification, 第10回計測自動制御学会制御部門マルチシンポジウム, 661-666, Mar. 2023. (Summary) This paper considers numerical analysis of state-term estimation in a closed-loop LPV (linear Parameter-Varying) identification method presented by the authors. Based on predictor-based system identification (PBSID) for an LTI (Linear-Time-Invariant) system, computation of the intersection between the pas and future data is motivated. The closed-loop LPV identification method presented by the authors is moreover reviewed, and the past and future part of the Gram matrices are derived. A computation method of the intersection between them is presented. (Keyword) LPV system / closed-loop idenitification / subspace identification / kernel method
5.	Kenji Ikeda and HIdeyuki Tanaka : Frequency Domain Uncertainty Estimation in a Closed Loop Subspace Identification Method, 第10回計測自動制御学会制御部門マルチシンポジウム, 667-674, Mar. 2023. (Summary) This paper proposes an estimate of the covariance matrix of the estimated system matrices in a closed-loop identification method previously proposed by the authors. The estimate of the covariance matrix can be derived from a set of input/output data used for the identification. It is important to provide a covariance matrix of the estimate in order to guarantee the quality of the estimate. Based on the estimated covariance matrix, the uncertainty of the estimated transfer function is derived. Numerical simulations illustrate the validity of the derived covariance estimate. (Keyword) Subspace identification / covariance analysis / frequency domain uncertainty
6.	Hideyuki Tanaka and Kenji Ikeda : Identification of Nonlinear ARX model using kernel method with random Fourier features, 第65回自動制御連合講演会, 935-940, Nov. 2022. (Summary) This paper studies identification of nonlinear-ARX (NARX) model using a kernel method. Random Fourier features are applied to identification of a NARX model, since the size of the matrix in computing regression becomes large in kernel methods, and the numerical computation time is long. Numerical simulation is given to evaluate effectiveness of the present method. (Keyword) NARX model identification / kernel method / Random Fourier features
7.	Kenji Ikeda and Hideyuki Tanaka : Covariance estimation of the estimated extended observability matrix in Closed-Loop MOESP, 第65回自動制御連合講演会, 922-927, Nov. 2022. (Summary) It is important to provide a covariance of the estimates to ensure the quality of the identification results. This paper proposes a formula of the covariance of the estimate of the extended observability matrix in Closed-Loop MOESP. The derived formula uses the gap between singular subspaces to estimate the perturbation of the extended observability matrix. The derived formula can be used to estimate its covariance matrix from a single identification result. Numerical simulation compares the sample covariance with the derived formula (Keyword) Subspace identification / Innovations model / Noise covarialnce / covariance of the estimate
8.	Hideyuki Tanaka and Kenji Ikeda : On state estimation of LPV system under feedback via kernel method, 第9回計測自動制御学会制御部門マルチシンポジウム, 701-708, Mar. 2022. (Summary) This paper studies state-space estimation for identifying a Linear Parameter Varying (LPV) system. By using kernel least-squares and kernel canonical correlation analysis, two state-space identification algorithms for a Linear-Time-Invariant (LTI) system are extended: The one is PBSID (Predictor Based System IDentification), and the other is the stochastic realization or stochastic subspace identification. It is pointed out in this paper that a closed-loop state-space identification algorithm for an LPV system recently developed coincides with the one obtained by extending the stochastic realization algorithm to an LPV system. Numerical simulations show that the algorithm based on stochastic realization suffers from numerical problems and that the algorithm based on PBSID outperforms that of stochastic realization. (Keyword) Linear Parameter Varying / State-Space Identification / PBSID / Stochastic realization / Kernel canonical correlation analysis
9.	Kenji Ikeda and HIdeyuki Tanaka : Numerical Analysis on the Rank of Weighting Matrix in BLUE, 第9回計測自動制御学会制御部門マルチシンポジウム, 45-52, Mar. 2022. (Summary) This paper numerically analyzes the rank of a weighting matrix in the BLUE (best linear unbiased estimator), which was previously proposed by the authors for the estimation of noise covariance matrix of the system. Since the weighting matrix is singular, its rank is required for the calculation of the pseudo-inverse. The rank of the weighting matrix is numerically analyzed and it is given as a function of the system degree, the number of output, and a design parameter called future horizon. (Keyword) Subspace identification / Gauss-Markov model / Noise covariance / Best linear unbiased estimator
10.	Hideyuki Tanaka and Kenji Ikeda : Idetification of minimal realization of plant under feedback, --- an approach via state estimation ---, 第64回自動制御連合後援会, 518-525, Nov. 2021. (Summary) This paper formulates a problem of finding a minimal realization of an unstable plant under feedback. Different from the earlier studies by the authors, the problem setting does not assume a linear-time-invariant controller of the finite order. The authors introduce an orthogonal direct sum decomposition of the input-output data in a Hilbert space and derive a state estimation method for a minimal realization of the plant based on the decomposition. They develop a realization algorithm for finding a minimal realization of an unstable plant based on the result and obtain a closed-loop subspace identification method. (Keyword) closed-loop identification / direct subspace identification / minimal realization / unstable plant / orthogonal projection
11.	Kenji Ikeda and Hideyuki Tanaka : On the covariance analysis of the estimated of the noise covariance, 第64回自動制御連合後援会, 526-533, Nov. 2021. (Summary) This paper analyzes the covariance of the estimated covariance matrix of the process and measurement noises in Gauss-Markov models. From some numerical simulations, it was shown that an estimation method of the noise covariance with CCA (canonical correlation analysis) weighting matrix proposed by the authors gives good performance. The covariance of the error term in the proposed method is derived. On the other hand, a BLUE (best linear unbiased estimate) for the corresponding problem is derived by using a newly derived lemma. The proposed estimate is compared with the BLUE. (Keyword) subspace identification / Innovations model / Noise covarialnce / Best linear unbiased estimate
12.	Hideyuki Tanaka and Kenji Ikeda : Minimal realization of unstable system under feedback, --- Derivation of numerically sound calculation method ---, 第8回計測自動制御学会制御部門マルチシンポジウム, 1050-1055, Mar. 2021. (Summary) This paper develops a numeirically sound method for identifying a minimal realization of an unstable linear plant under feedback. It is assumed that the plant and noise models have non-common state matrices and that the plant is stabilized by a linear time invariant system of finite order. A numerically sound computation method is derived by the product of the extended observability and reachability matrices of the left and right coprime factor representations of the plant, by improving the past research of the authors. Numerical simulation results show the effectiveness of the proposed method. (Keyword) close-loop identification / minimal realization / unstable plant / joit input-ouput approach
13.	Kenji Ikeda and HIdeyuki Tanaka : Numerical Analysis of the Reduction of the Bias in Closed-Loop Subspace Identification Method, 第8回計測自動制御学会制御部門マルチシンポジウム, 1044-1049, Mar. 2021. (Summary) It is seen that the estimate of the extended observability matrix in Closed-Loop MOESP has a smaller bias compared to other methods when pas/future horizons are comparatively small. The reason of the fact is that the image of the term causing the bias is close to the image of the extended observer matrix. In this paper, it is analyzed theoretically and numerically how close the images are and numerical simulation shows that the bias is reduced in practice. (Keyword) Subspace identification / Innovations model / Unstable systems
14.	Yamano Yuya and Kenji Ikeda : Selection of Input Joints for Gait Generation of Biped Robot, 第7回制御部門マルチシンポジウム, 597-600, Mar. 2020. (Summary) エネルギー制御を用いて二足歩行ロボットの歩容生成をするとき，加えるエネルギーが同じでも入力する関節やタイミングが異なると異なる歩容が生成される．入力トルクによってつぎ込むエネルギー量が変わる．そのためトルクを入力する関節を選択しなければならない．本研究では，効率の良い歩容生成を目的とし，エネルギー制御によって求めたトルクを入力する関節の選択について考察した．検証として，腰に入力した場合と膝に入力した場合を比較した．その結果，エネルギー効率の良い入力関節が膝であることがわかった． (Keyword) system identification / Unscented transformation / robust nonlinear estimation
15.	Matsuo Kotaro and Kenji Ikeda : Estimation of parameters using system identification and application to robust nonlinear estimation, 第7回制御部門マルチシンポジウム, 523-526, Mar. 2020. (Summary) In this study, we propose an estimation method of physical parameters of the plants and their uncertainty of the controlled object using the system identification and Unscented Transformation (UT). By accurately estimating the uncertainty, the deterioration of the precision of the state estimate is expected to be reduced. (Keyword) system identification / Unscented transformation / robust nonlinear estimation
16.	Daikoku Takahiro and Kenji Ikeda : Modeling of Drone's Energy Consumption, 第62回自動制御連合講演会, 1-4, Nov. 2019. (Summary) In this paper, we focus on drone's battery exhaustion. The purpose is to extend the flight time by controlling the flight speed. We obtain electric current data using amperometer on the drone. Next, we estimate a model of energy consumption using system identification an compare the energy consumption of the real machine and the estimated model. (Keyword) subspace identification / innovations model / semi-definite programming
17.	Kenji Ikeda and Hideyuki Tanaka : Identification of systems having unreachable stochastic system with eigenvalues on the closed unit disc, 第6回制御部門マルチシンポジウム, 342-347, Mar. 2019. (Summary) This paper applies an identification method of innovations model previously proposed by the authors to the case when the system is unstabilizable. It is shown that the previously proposed optimization problem can estimate the noise covariance even when the system is unstabilizable. The step of estimating the innovations model from the noise covariance is modified so as to treat the unstabilizable systems. A numerical simulation illustrates the effectiveness of the proposed method. (Keyword) subspace identification / innovations model / unstabilizable systems
18.	Kenji Ikeda and Hideyuki Tanaka : Another semidefinite programming problem for subspace identification method taking initial state into account, 第61回自動制御連合講演会, 205-212, Nov. 2018. (Summary) This paper proposes an another semi-definite programming problem for the consistent estimation of the noise covariance and analyzes the solvability of the problem. The authors proposed two methods: one is to estimate the noise covariance together with the bias term, and the other is to estimate the noise covariance by cancelling the bias term. This presentation concerns the latter approach. The merit of the proposed method is the size of the data matrix required in the proposed method is smaller than that in the previous method. Conditions for the problem to be solved and the uniqueness of the spectral density function of the noise are analyzed. (Keyword) subspace identification / innovations model / semi-definite programming
19.	Kenji Ikeda and Hideyuki Tanaka : On the Identifiability in an SDP Problem for the Estimation of Innovations Model, 第5回制御部門マルチシンポジウム, 1183-1188, Mar. 2018. (Summary) The authors proposed an SDP formulation for the identification of the Kalman gain and the covariance matrix of the innovations process, aiming at a consistent estimation. However, the rank of a coefficient matrix of the variables does not have a column full rank, so that the uniqueness of the solution or the identifiability are not clear. In this paper, a sufficient condition for the system to be identifiable by the proposed method almost everywhere is derived.
20.	田中秀幸 and Kenji Ikeda : Stochastic subspace identification methods taking initial state into account via semidefinite programming approach, 第60回自動制御連合講演会, 1447-1454, Nov. 2017. (Summary) In this paper, the authors develop stochastic subspace identification algorithms taking initial state into account via a semidefinite programming approach. They develop two subspace algorithms by deleting the initial state from the covariance matrix and by finding the white-input and the state terms in the covariance data. They moreover consider variations of the computation algorithms based on the minimum and maximum convariance matrix of the state of the stochastic system. They provide numerical simulation, comparing the algorithms with the one previously proposed. (Keyword) stochastic subspace identification / stochastic realization / initial state / Kalman gain
21.	Kenji Ikeda and Hideyuki Tanaka : Problem Formulation for Estimating Innovations Model, 第60回自動制御連合講演会, 1441-1446, Nov. 2017. (Summary) An estimation problem of kalman gain and the covariance of the innovations process has been formulated as a nonlinear optimization problem in order to reduce the bias of the estimate. In this paper, the problem is formulated as an SDP (semi-definite programming) problem by using change of variables. The effectiveness of the proposed method is illustrated by using a numerical example. (Keyword) subspace identification / consistent estimate / innovations model / linear matrix inequality
22.	Hideyuki Tanaka and Kenji Ikeda : Closed-loop identification based on the null-space of the extended observability matrix, The 4th Multi-symposium on Control Systems (MSCS 2017), 1357-1364, Mar. 2017. (Summary) In this paper, the authors consider closed-loop identification based on a stochastic system for the joint-input-output. They develop a new closed-loop system identification method based on the null-space of the extended observability matrix for the stochastic system by taking the initial state of the stochastic system into account. Numerical simulation indicates that the new method gives better results than the closed-loop system identification method developed earlier by the authors based on a Riccati equation and linear matrix inequalities to guarantee positive realness of the stochastic system. (Keyword) closed-loop system identification / stochastic realization / innovation representation / null-space / extended observability matrix
23.	Kenji Ikeda : Local Convergence of an Algorithm for Consistent Estimate of Innovations Model, 第4回制御部門マルチシンポジウム, 1365-1370, Mar. 2017. (Summary) This paper analyzes the convergence property of a numerical algorithm proposed by the author, which estimates Kalman gain and the covariance matrix of a innovations process consistently. The estimates are given as solutions of some BMI problem, so an algorithm which iteratively solves some LMI's are proposed. In this paper, the algorithm is shown to be locally convergent.
24.	Kenji Ikeda : Convergence of an Estimation Method of kalman Gain by using Cholesky Factor, 第59回自動制御連合講演会, 455-458, Nov. 2016.
25.	Kenji Ikeda : Consistent Estimate of Kalman Gain by using Cholesky Factor, 第60回システム制御情報学会研究発表講演会, 1-4, May 2016.
26.	Kenji Ikeda : Numerical Solution for a Consistent Estimate of kalman Gain, 第3回制御部門マルチシンポジウム, 403-408, Mar. 2016.
27.	Kenji Ikeda : Consistent Estimate of Covariance Matrix of Innovations Process, 第58回自動制御連合講演会, 1-6, Nov. 2015. (Keyword) system identification / subspace identification / innovations process / error analysis
28.	Kenji Ikeda : Error Analysis of Subspace Identification Method in Closed Loop Environment, 第59回システム制御情報学会研究発表講演会, 1-4, May 2015. (Summary) This paper analyzes the estimation error of B and D matrices in closed-loop environment which are estimated by using OE-PIV method proposed by Chou and Verhaegen. Estimates of A and C matrices in closed-loop environment which are obtained by applying PO-MOESP method in the framework of direct method of the closed loop identification has no asymptotic bias and their estimation error can be derived in the previous paper. In this paper, the relation between the estimation error of (B,D) and the estimation error of (A,C,K) are derived.
29.	Kenji Ikeda : Error Analysis of PO-MOESP method in Closed Loop Environment, --- Analysis on the estimates of A and C matrices ---, The 2nd Multi-symposium on Control Systems, 55-60, Mar. 2015. (Keyword) system identification / Subspace mode identification / error analysis / closed loop systems
30.	Kenji Ikeda : Error Analysis of PO-MOESP method in Closed Loop Environment, --- Analysis of the estimates of A and C matrices ---, 第2回制御部門マルチシンポジウム, 55-60, Mar. 2015. (Summary) This paper analyzes the estimation error of A and C matrices in PO-MOESP method in closed loop environment. The estimates of B and D in PO-MOESP method will be biased in closed loop environment while the estimates of A and C will be unbiased. So, PO-MOESP method is still applicable for the estimation of A and C matrices in closed loop environment. In the previous analysis, the plant is assumed to be stable, and there are error terms which can be neglected when the future horizon is set to be large. In the closed loop environment, the plant might be unstable and the term will not be negligible. In this paper, the error terms are analyzed in detail and their explicit formulations are derived. (Keyword) System Identification / Subspace model identification / Error Analysis / Closed Loop Systems
31.	Kenji Ikeda and HIroshi Oku : Error Analysis of B and D matrices in PO-MOESP Method, 第57回自動制御連合講演会, 1416-1419, Nov. 2014. (Summary) Estimation errors of B and D matrices in PO-MOESP method are analyzed. The gap related amount plays an important role in the analysis of the estimation errors. Not only the most dominant terms but also the orders of the subordinate terms of the estimation errors are analyzed. (Keyword) Modeling / Subspace Identification / Error Analysis
32.	Kenji Ikeda : Comparison of Shift Invariance/State Approaches in Subspace Identification Methods, 第58回システム制御情報学会研究発表講演会, 1-6, May 2014. (Keyword) system identification / analysis of variance / Subspace model identification
33.	Kenji Ikeda : Error Analysis of PO-MOESP method, 第1回制御部門マルチシンポジウム, 1108-1113, Mar. 2014. (Keyword) system identification / subspace model identification / error analysis
34.	Kenji Ikeda : OE/ARMAX, 第56回自動制御連合講演会, 483-486, Nov. 2013. (Summary) Identification method based on ARMAX model gives better performance when the system is contaminated by a colored noise because noise model as well as plant model is estimated. However, in order to obtain an plant model, some process such as model reduction or decoupling will be required. Thus, the influence of such process to the precision of the plant model must be analyzed. This paper analyzes the estimation error in $A$ matrix of the plant model when OE model and ARMAX are adopted. (Keyword) subspace model identification / output error model / ARMAX model / variance analysis
35.	松村剛, Kenji Ikeda and Takao Shimomura : Energy-recovery control of bipedal walking robot, 2013年度計測自動制御学会四国支部学術講演会, 175-176, Nov. 2013. (Summary) 受動歩行は, 人間の歩行のように自然で滑らかな歩行であるため, エネルギー効率に優れていると考えられ注目されている. しかし, 受動歩行を行うと二足歩行ロボットは膝が伸びきったり, 地面に脚をついたときに運動エネルギーが減少している. このエネルギーを回生し, さらにエネルギー効率の良い歩行を実現する.
36.	中野秀信, Kenji Ikeda and Takao Shimomura : Stabilization of the bipedal walking robot using orthogonal function expansion, 2013年度計測自動制御学会四国支部学術講演会, 123-124, Nov. 2013. (Summary) 受動歩行を行う二足歩行ロボットは制御を必要としない代わりに小さい外乱で倒れてしまう. そこで, 非線形システムを周期システムとみなし, 直交関数展開を用いて制御を行うことで倒れやすいパターンでの転倒を防ぐ.
37.	加藤渓祐, Kenji Ikeda and Takao Shimomura : Web アプリケーションのデバッグ・保守支援, 電気関係学会四国支部連合大会, Vol.17, No.5, 337, Sep. 2013.
38.	Kenji Ikeda : Variance analysis of plant model using MOESP type methods, 第57回システム制御情報学会研究発表講演会, May 2013. (Summary) Precision of the estimated plant model is often the main interest of system identification. In order to take the predictable part of the noise into account, the noise model is estimated together with the plant model by using ARMAX model. In that case, a model reduction procedure will be required in order to obtain the plant model. On the other hand, the plant model can be estimated directly by using output error (OE) model. In this paper, PI-MOESP method and PO-MOESP method are compared by analysing the signal and noise components of the estimated plant model under the assumption that there are no common poles in the plant and the noise models. The magnitude of the noise component in each method is discussed when the past or future horizon varies and it is shown that there is a possibility that PI-MOESP method gives better performance than PO-MOESP method. A numerical example is given to illustrate the discussions. (Keyword) subspace model identification / variance analysis
39.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Linear Approximation Model for Bipeds Robot and its Application, SICE 13th Annual Conference on Control Systems, Mar. 2013. (Summary) Modeling method of a biped robot by using periodic property is proposed. The input and the output of the robot is sampled by using lifting technique and discretized by expanding the sampled signals by using orthonal functions. The coefficients of the sampled input signal are restricted to be of finite dimension while the coeeficients of the output signal are projected to a finite dimensional set. In order to estimate the linear approximation model, a subspace identification method is utilized. (Keyword) biped robot / periodic system / subspace identification method
40.	Kenji Ikeda : Asymptotic Variance Analysis of Bias Compensated State Space Model Identification Method, 第55回自動制御連合講演会, 32-35, Nov. 2012. (Summary) Based on the analysis of asymptotic variance in bias compensated state space model identification(BCSS) method, PI-MOESP method and PO-MOESP method are compared under the condition that there are no common poles in plant model and noise model. It is shown that there is a possibility that PI-MOESP method gives better performance than PO-MOESP and a numerical example is given to show such case. (Keyword) estimation / バイアス補償 / analysis of variance
41.	西村拓也, Kenji Ikeda and Takao Shimomura : サーバベースWeb アプリケーション回帰統合テスト支援方式, 電気関係学会四国支部連合大会, 327, Sep. 2012.
42.	木村心, Kenji Ikeda and Takao Shimomura : Waiting ブロック・ツリー型デッドロックの検出, 電気関係学会四国支部連合大会, 290, Sep. 2012.
43.	Takahiro Horibe, Yoshio Mogami, Kenji Ikeda, Norihiko Ono and Yasuyuki Ibaraki : Construction of Stock Trading Support System Using Neural Network, 第56回システム制御情報学会研究発表講演会講演論文集, 373-374, May 2012.
44.	Kazuiro Koguro, Yoshio Mogami, Kenji Ikeda and Norihiko Ono : Forecasting System of Actual Temperature and Humidity Using Stochastic Real-Valued Unit, 第56回システム制御情報学会研究発表講演会講演論文集, 371-372, May 2012.
45.	Yuta Kanzaki, Yoshio Mogami, Kenji Ikeda and Norihiko Ono : Autonomous Transfer System in Environment Which Have Irregular Moving Obstacles, 第56回システム制御情報学会研究発表講演会講演論文集, 173-174, May 2012.
46.	Tomoaki Kaihara, Yoshio Mogami, Kenji Ikeda and Norihiko Ono : Autonomous Cooperation System of Mobile Objects Using Stochastic Real-Valued Unit, 第56回システム制御情報学会研究発表講演会講演論文集, 171-172, May 2012.
47.	Kenji Ikeda and Hiroshi Oku : Asymptotic Properties of Bias Compensated State Space Model Identification Method in Closed Loop Environment, SICE 12th Annual Conference on Control Systems, Mar. 2012. (Summary) Bias compensated state space model identification method in close loop environment (CL-BCSS) is introduced. The noise is assumed to be a 0 mean colored noise. It is also assumed that the reference input is a white Gaussian process uncorrelated to the noise. Signal and noise components of a certain matrix in the proposed method and CL-MOESP method are anlyzed and compared. It is shown that CL-BCSS method corresponds to CL-MOESP method with infinite past horizon. (Keyword) closed-loop identification / state-space model identification / bias compensation / variance analysis
48.	Kenji Takatori, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Development of a victim search system by using autonomous learning unit, 第20回計測自動制御学会中国支部学術講演会論文集, 230-231, Nov. 2011.
49.	Yasuyuki Ibaraki, Yoshio Mogami, Norihiko Ono, Kenji Ikeda and Syunji Nozumi : ニューラルネットワークを用いた株価短期変動レンジ予測システムの構築, 2011年度計測自動制御学会四国支部学術講演会論文集, Nov. 2011.
50.	Tomoaki Tamada, Yoshio Mogami, Norihiko Ono, Kenji Ikeda and Haruhiko Tomida : スイッチング非定常環境下での学習ユニットによる未知大域的最適解探索, 2011年度計測自動制御学会四国支部学術講演会論文集, Nov. 2011.
51.	Kenji Ikeda : Variance Analysis of Bias Compensated State Space Model Identification Method, --- Case of 1st order system ---, 40th SICE Symposium on Control Theory, 153-158, Sep. 2011. (Summary) A method of bias compensation in subspace identification method is introduced and an analysis on its variance is presented. The noise is assumed to be a 0 mean colored noise and is assumed to be uncorrelated with the input. The variance of the extended observability matrix is analyzed and a closed form of the variance is derived for the case when the plant is a 1st order system and the noise is white. (Keyword) State Space Identification / Variance Analysis / Bias Compensation
52.	石田達大, Kenji Ikeda and Takao Shimomura : An Approach to Design Pattern-Based Refactoring, 電気関係学会四国支部連合大会, 278, Sep. 2011.
53.	Ryuichi Yuasa, Yoshio Mogami, Norihiko Ono, Kenji Ikeda, Masahiro Miyamichi and Hiroki Uesugi : Construction of Autonomous Sailing System of Ship with SRV Learning Unit, Proceedings of the 55th Annual Conference of the Institute of Systems, Control and Information Engineers, 455-456, May 2011.
54.	Masaki Sugikawa, Yoshio Mogami, Norihiko Ono, Kenji Ikeda, Masahiro Miyamichi and Hiroki Uesugi : Forecasting System of Actual Sunshine Duration based on Weather Statistics Information, Proceedings of the 55th Annual Conference of the Institute of Systems, Control and Information Engineers, 453-454, May 2011.
55.	Yasuyuki Ibaraki, Yoshio Mogami, Norihiko Ono, Kenji Ikeda, Natsuki Oba and Syunji Nozumi : Short-term Change Prediction of Stock Prices by Using Neural Network, Proceedings of the 55th Annual Conference of the Institute of Systems, Control and Information Engineers, 451-452, May 2011.
56.	Tomoaki Tamada, Yoshio Mogami, Norihiko Ono, Kenji Ikeda, Hiroki Okada and Haruhiko Tomida : Global Optimal Solution Search under Unknown Switching Environment, Proceedings of the 55th Annual Conference of the Institute of Systems, Control and Information Engineers, 403-404, May 2011.
57.	Kenji Ikeda : Variance Analysis of Bias Compensated State Space Model Identification Method, SICE 11th Annual Conference on Control Systems, Mar. 2011. (Summary) A method of bias compensation in subspace identification method is introduced and an analysis on its variance is presented. The noise is assumed to be a 0 mean colored noise and is assumed to be uncorrelated with the input. The variance is analyzed from a view point of S/N ratio of a matrix before the singular value decomposition and then perturbation on the left singular subspace is analyzed. (Keyword) state space identification / Variance Analysis / Bias Compensation
58.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-Compensated State Space Identification Method, --- A Numerical Analysis of the Gap between the Singular Subspaces ---, 第11回適応学習制御シンポジウム資料, 61-66, Jan. 2011. (Summary) A method of bias compensation in subspace identification method is introduced and a numerical analysis of its variance is presented. The noise is assumed to be a 0 mean colored noise and is assumed to be uncorrelated with the input. The variance is analyzed from a view point of S/N ratio of a matrix before the singular value decomposition. However, the numerical analysis reveals that the gap between the nominal and perturbed singular subspaces is slightly different from the S/N ratio of the matrix. (Keyword) state space identification / variance analysis / bias compensation
59.	Shunji Nozumi, Yoshio Mogami, Norihiko Ono, Kenji Ikeda and Natsuki Ooba : Stock buying and selling prediction using neural network and stock automatic buying and selling system, 2010年度計測自動制御学会四国支部学術講演会論文集, Nov. 2010.
60.	Haruhiko Tomida, Yoshio Mogami, Norihiko Ono, Kenji Ikeda and Hiroki Okada : Construction of Stochastic Estimator-Type Hierarchical Structure Learning Algorithm and Its Preformance, 2010年度計測自動制御学会四国支部学術講演会論文集, Nov. 2010.
61.	Kenji Takatori, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Construction of Autonomous Transfer System for Rescue by Plural Movement Bodies, 2010年度計測自動制御学会四国支部学術講演会論文集, Nov. 2010.
62.	Hiroki Uesugi, Yoshio Mogami, Norihiko Ono, Kenji Ikeda and Masahiro Miyamichi : Property of Stochastic Real-Valued Unit in Noisy Environment, 2010年度計測自動制御学会四国支部学術講演会論文集, Nov. 2010.
63.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Numerical Analysis of Bias Compensated State Space Model Identification, 第53回自動制御連合講演会論文集, 678-683, Nov. 2010. (Keyword) close-loop identification / state-space mmethods / bias compensation / variance
64.	Kenji Ikeda, Tomoaki Kashiwao and Takao Shimomura : Estimation of a Transition Time in Piecewise Linear Systems, 第10回計測自動制御学会制御部門大会資料, Mar. 2010.
65.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-Compensated State Space Identification Method, 第10回適応学習制御シンポジウム資料, 91-94, Jan. 2010.
66.	Tomoaki Kashiwao, Kenji Ikeda and Takao Shimomura : Estimation of the transition time of piecewise linear systems by using wavelet analysis, 第52回自動制御連合講演会講演論文集, Nov. 2009.
67.	Natsuki Oba, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Prediction of Stock Prices Change by Neural Network, 第52回自動制御連合講演会講演論文集, Nov. 2009.
68.	Ryo Nishitani, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Consturuction and verification of stock prices forecast system using neural networks, 第52回自動制御連合講演会講演論文集, Nov. 2009.
69.	Hiroki Okada, Yoshio Mogami, Norihiko Ono, Norio Baba and Kenji Ikeda : Application of Continuous Action Learning Automata in Global Optimal Solution Search of Unknown Function, 第52回自動制御連合講演会講演論文集, Nov. 2009.
70.	Masahiro Miyamichi, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Construction of Autonomous Transfer System by Plural Movement Bodies Using Stochastic Real-Valued Unit, 第52回自動制御連合講演会講演論文集, Nov. 2009. (Keyword) stochastic real-valued unit / autonomous transfer system / Moving Objects / intruder capture problem / stochastic learning
71.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : A parameter estimation method based on a reduced order plant model, 38th SICE Symposium on Control Theory, 1-4, Sep. 2009. (Keyword) parameter estimation / reduced order model
72.	Natsuki Oba, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Change Prediction of Stock Prices by Using Neural Network, Proceedings of the 53rd Annual Conference of the Institute of Systems, Control and Information Engineers, 631-632, May 2009.
73.	Masahiro Miyamichi, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Intruder Capture System by Autonomous Transfer System Using Stochastic Real-Valued Unit, Proceedings of the 53rd Annual Conference of the Institute of Systems, Control and Information Engineers, 497-498, May 2009.
74.	Hiroki Okada, Yoshio Mogami, Norihiko Ono, Norio Baba and Kenji Ikeda : Global Optimal Solution Search of Unknown Function Using Continuous Action Learning Automata, Proceedings of the 53rd Annual Conference of the Institute of Systems, Control and Information Engineers, 465-466, May 2009.
75.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-compensated least squares method in closed loop environment, --- Identification under colored noise ---, Proceedings of SICE 9th Annual Conference on Control Systems, Mar. 2009. (Keyword) closed loop identification / least squares method / bias compensation
76.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-Time Model Estimation in Closed-Loop Environment, --- On the relaxation of the identifiability condition ---, 9th SICE Symposium on Adaptive and Learning Control, 1-4, Jan. 2009.
77.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias Compensated Recursive Least Squares Estimate in Closed Loop Environment, 第51回自動制御連合講演会, 584-587, Nov. 2008. (Keyword) recursive least squares estimate / bias compensation / closed loop identification
78.	赤峰利幸, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Autonomous Transfer System for Mobile Object Based on Stochastic Learning, 2008 SHIKOKU-SECTION JOINT CONVENTION RECORD OF THE INSTITUTES OF ELECTRICAL AND RELATED ENGINEERS, 321, Sep. 2008.
79.	高畠真大, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Forecast System of Stock Prices by Nueral Networks, 2008 SHIKOKU-SECTION JOINT CONVENTION RECORD OF THE INSTITUTES OF ELECTRICAL AND RELATED ENGINEERS, 320, Sep. 2008.
80.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias Compensated Recursive Least Squares Estimate in Closed Loop Environment, --- On the improvement of the convergency ---, 37th SICE Symposium on Control Theory, 105-110, Sep. 2008.
81.	Yusuke Shimada, Yoshio Mogami, Norihiko Ono, Norio Baba and Kenji Ikeda : Construction of pursuit-type variable hierarchical structure learning algorithm and its performance, Proceedings of the 52nd Annual Conference of the Institute of Systems, Control and Information Engineers, 567-568, May 2008.
82.	Koichiro Murata, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Autonomous transfer system for mobile object using stochastic real-valued learning unit, Proceedings of the 52nd Annual Conference of the Institute of Systems, Control and Information Engineers, 565-566, May 2008.
83.	Mami Sawada, Kenji Ikeda, Takao Shimomura and Yoshio Mogami : Continuous-time model estimation in closed loop environment, --- Iterative algorithm by using bias-compensated least squares method ---, Proc. of the 52nd Annual Conference of ISCIE, 331-332, May 2008. (Summary) Continuous-time model estimation in closed loop environment is proposed. The asymptotic bias of least squares estimate may be made zero by designing the prefilter appropriately. However in general, iterative re-design of the prefilter is required because it depends on the plant parameters. Bias-compensated least squares method is considered to improve the asymptotic bias of the LS estimate and it is considered to improve the convergence speed of the iteration. In this paper, a continuous-time model estimation method in closed loop environment by using bias-compensation is proposed.
84.	Ryo Nishitani, Yoshio Mogami, Norihiko Ono and Kenji Ikeda : Consturuction of stock prices forecast system by using neural networks, Proceedings of the 52nd Annual Conference of the Institute of Systems, Control and Information Engineers, 507-508, May 2008.
85.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-compensated Least Squares Method in Closed Loop Environment, --- Fast iterative algorithm ---, 第8回制御部門大会, Mar. 2008. (Keyword) Closed Loop Identification / Bias compensated least squares method
86.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-Compensated Least Squares Method in Closed Loop Environment, --- A filter design with colored noise ---, 8, 57-60, Jan. 2008. (Summary) In this paper, a bias-compensated least squares(BCLS) method in the closed loop environment is extended for the cases where the measurement noise is colored. Namely, a BCLS method in the closed loop environment based on Box-Jenkins model is proposed. Noise model is estimated based on a non-parametric mode or high order model because the purpose of the closed loop identification is th eimprovement of the plant dynamics and the purpose of the identification of the noise model is to obtain the plant parameter with smaller variances.
87.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-compensated Least Squares Method in Closed Loop Environment, --- A formulation based on the transfer function representation ---, 第50回自動制御連合講演会, 320-323, Nov. 2007. (Keyword) parameter estimation / least squares method / closed loop systems / bias compensation
88.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Bias-compensated Least Squares Method in Closed Loop Environment, 第36回制御理論シンポジウム資料, 143-146, Sep. 2007. (Summary) In this paper, a bias-compensated least squares method in the closed loop environment is proposed. It is assumed that the observation noise is a white gaussian signal while there are no process noises. It is also assumed that the plant is controlled by a linear time invariant controller and that the closed loop system is asymptotically stable. It is shown that the asymptotic bias of the least squares estimate in the closed loop environment is represented as a difference between the parameter of the prefilter and that of an "optimal prefilter" and that the proposed bias-compensated least squares estimate is consistent. (Keyword) parameter estimation / least squares method / closed loop systems / bias compensation
89.	Mami Sawada, Kenji Ikeda and Takao Shimomura : Continuous-time model estimation by using adaptive observer, --- Consistency of estimated parameters ---, Proceedings of the 51st Annual Conference of the Institute of Systems, Control and Information Engineers, 235-236, May 2007. (Keyword) continuous-time model / estimation / consistency
90.	Masayuki Imoto, Yoshio Mogami, Norihiko Ono, Norio Baba, Kenji Ikeda and Takao Shimomura : Search for Global Optimal Solution of The Unknown Function by Using Variable Hierarchical Structure Learning Automata, 講演論文集, 717-718, May 2007.
91.	Toshiyuki Akamine, Yoshio Mogami, Norihiko Ono, Kenji Ikeda and Takao Shimomura : Autonomous Transfer System for Mobile Object Using Stochastic Real-Valued Reinforcement Learning Unit, 講演論文集, 681-682, May 2007.
92.	Masahiro Takahata, Yoshio Mogami, Norihiko Ono, Norio Baba, Kenji Ikeda and Takao Shimomura : Stock prices forecast system by using neural networks, 講演論文集, 637-638, May 2007.
93.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-time model estimation by using bias compensated adaptive observer, 第7回適応学習制御シンポジウム資料, 93-98, Jan. 2007. (Keyword) bias compensation / continuous-time model / adaptive observer
94.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-time model estimation method by using adapitve observer, --- Iterative design of the state variable filters ---, 第49回自動制御連合講演会資料, Nov. 2006. (Keyword) continuous-time model / estimation / adapitve observer / state variable filters / Kalman filter
95.	Makio Kinoshita, Yoshio Mogami, Kenji Ikeda and Takao Shimomura : The area weather forecast system based on improvement-type reinforcement learning unit, 第49回自動制御連合講演会, Nov. 2006.
96.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-time model identification by using adaptive observer, --- Redesign of state variable filters ---, Proc. of SICE 6th Annual Conference on Control Systems, 25-28, May 2006. (Summary) This paper proposes a redesign method of the state variable filters(SVF's) which is used in the previously proposed continuous-time model estimation method by using adaptive observer. The SVF's are designed by using a Kalman filter which minimizes the variance of the residuals. (Keyword) continuous-time model / sampled-data systems / adaptive filters / Kalman filter
97.	Kazumasa Sakao, Kenji Ikeda and Takao Shimomura : Parameter estimation of a continuous-time nonlinear system by using unscented Kalman filter, 第50回システム制御情報学会研究発表講演会講演論文集, 291-292, May 2006. (Summary) This paper proposes a continuous-time parameter estimation algorithm for nonlinear systems by using the Unscented Kalman Filter(UKF). The Ukf algorithm extends the Kalman Filter to nonlinear systems by transforming approximations of the probability distributions through the nonlinear process and measurement functions. Therefore, it is possible to calculate the estimate based on continuous-time model by using numerical integration, and make better estimation than the Extended Kalman Filter(EKF). (Keyword) Unscented Transformation / Kalman Filter / Continuous-time model / parameter estimation
98.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-time model identification by using adaptive observer, --- Estimation of the noise model ---, Proc. of 6th SICE Symposium on Adaptive and Learning Control, 31-34, Mar. 2006. (Summary) This paper deasl with the estimation of the noise model in the previously proposed continuous-time model estimation method by using adaptive observer. Hybrid Box-Jenkins model is adopted for the noise model. (Keyword) Continuout-time model / adapitve observer / noise model
99.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Stability analysis of continuous-time model identification method by using adaptive observer, 第48回自動制御連合講演会資料, 639-642, Nov. 2005. (Summary) This paper proposes a continous-time model estimation method from sampled I/O data with the ZOH input assumption. State variable filters are modified in order to estimate the intersample signals by using a previously obtained parameter estimate. The bias of the estimated parameters is analyzed. (Keyword) continuous-time model / system identification / adaptive observer
100.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-time model identification by using adaptive observer, --- On the existence of an optimal solution ---, Proc. of the 34th SICE Symposium on Control Theory, 133-136, Oct. 2005. (Summary) This paper proposes a continuous-time model estimation method from sampled I/O data with the ZOH input assumption. State variable filters are modified in order to estimate the intersample signals by using a previously obtained parameter estimate. A sufficient condition for the parameter convergence is analyzed. (Keyword) continuous-time model / adaptive observer / state variable filter / contraction mapping theorem
101.	Yoshio Mogami, Mitsuhiro Matsumoto, Takao Shimomura and Kenji Ikeda : Application of the variable hierarchical structure learning autmata to the search for global optimal solution of the unknown multimodal objective function, 平成17年度電気関係学会四国支部連合大会講演論文集, 305, Sep. 2005.
102.	Yoshio Mogami, Yuki Miyaura, Kenji Ikeda and Takao Shimomura : The area weather forecast system using reinforcement learning unit, 平成17年度電気関係学会四国支部連合大会講演論文集, 304, Sep. 2005.
103.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : A design of state variable filters in continuous-time model identification, Proc. of SICE 5th Annual Conference on Control Systems, 49-52, May 2005. (Summary) The authors have proposed an identification method of continuous-time models by using adaptive observers. However, it was not clear how to construct the state variable filters, which will affect the performance of the identification. In this paper, several kinds of the realization of the state variable filters are compared by means of numerical simulations. (Keyword) continuous-time model / adaptive observer / state variable filters
104.	Makio Kinoshita, Yoshio Mogami, Takao Shimomura and Kenji Ikeda : Regional rainfall forecast system based on reinforcement learning, 第49回システム制御情報学会研究発表講演会講演論文集, 291-292, May 2005.
105.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-Time Model Identification By Using Adaptive Observer, --- Generalization of the State Variable Filters ---, 第47回自動制御連合講演会, Nov. 2004.
106.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : A recursive algorithm of continuous-time model identification for MIMO systems, 33rd SICE Symposium on Control Theory, 229-232, Nov. 2004.
107.	坂尾一将, Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Accuracy Improvement of Subspace Model Identification of Continuous Time System, 平成16年電気関係学会四国支部連合大会講演論文集, 80, Sep. 2004.
108.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : Continuous-time model identification by using fast adaptive observer, 第4回制御部門大会資料, 231-236, May 2004. (Summary) This paper introduces a new method to identify a continuous-time model from sampled I/O data by using an adaptive observer based on RLS(Recursive Least Squares) method. The convergence of the parameter estimate is considereably improved compared to the steepest descent method. In order to identify the plant from a finite number of I/O data, an adaptive observer for a backward system is also proposed. (Keyword) Continuous-time model / identification / Sampled-data system / adaptive observer
109.	Mitsuhiro Matsumoto, Yoshio Mogami, Norio Baba, Takao Shimomura and Kenji Ikeda : Search for Global Optimal Solution of The Unknown Function by Using Variable Hierarchical Structure Learning Automata, Proceedings of the 48th Annual Conference of the Institute of Systems, Control and Information Engineers, 35-36, May 2004.
110.	Yuuki Miyaura, Yoshio Mogami, Kenji Ikeda and Takao Shimomura : The Area Weather Forecast System Using Reinforcement Learning Unit, Proceedings of the 48th Annual Conference of the Institute of Systems, Control and Information Engineers, 461-462, May 2004.
111.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : 適応観測器を用いた連続時間モデル同定, 第32回制御理論シンポジウム資料, 261-264, Nov. 2003.
112.	Takao Shimomura, 高橋宗雄, Kenji Ikeda and Yoshio Mogami : 動的テーブルによるWebアプリケーション自動合成, 第2回情報科学技術フォーラム, 181-182, Sep. 2003. (Summary) ソフトウェアはプログラム言語を用いて記述されるが，記述されたプログラムコードを直感的に理解するのは容易ではない．このため，グラフィックスを用いた，さまざまのソフトウェア開発支援技術が研究されてきた．本論文では，Webアプリケーションの開発を例にとり，グラフィックスを用いることにより，ユーザが頭の中で描いたイメージに従ってソフトウェアを開発することができるイメージ指向プログラミング技法を提案し，データベースやプログラムテーブルの内容に基づいて，様々な種類のWebテーブルを動的に構成するWebアプリケーション自動合成方式について述べる．
113.	Kenji Ikeda : サンプル値系の適応観測器, 第31回制御理論シンポジウム資料, 235-238, Oct. 2002.

Et cetera, Workshop:

1.	MICHIRU Katayama, Kenji Ikeda and Tetsushi Ueta : Laminar stabilization control in chaos intermittency, IEICE Technical Report, Vol.NLP2023, No.28, 65-68, Jun. 2023. (Summary) Intermittency gives the irregular switching of almost periodic (laminar) and chaotic (burst) motions, and can be confirmed in every physical system within a certain parameter range. In this paper, by utilizing a method stabilizing a periodic motion which have been vanished by a saddle node bifurcation, a laminar motion in the intermittency is stabilized by a tiny control input. A pseudo periodic orbit composed by the vanished orbit or laminar motion is applied as a reference for the linear feedback control. We demonstrate the control for a circuit model and show robust stabilization responses. (Keyword) chaos control / chaos intermittency
2.	Kenji Ikeda, Yoshio Mogami and Takao Shimomura : 適応観測器を用いた連続時間モデル同定における収束条件について, 第4回適応学習制御シンポジウム資料, 111-114, Mar. 2004.
3.	Takao Shimomura, Muneo Takahashi, Kenji Ikeda and Yoshio Mogami : Web アプリケーションにおけるイメージ指向プログラミングの試み, 情報処理学会プログラミング研究会発表資料, 1-7, Jan. 2003. (Summary) Webアプリケーションの開発において，開発者が頭の中で描いたイメージに従ってソフトウェアを開発することを可能とする，グラフィックスを用いたイメージ指向プログラミング技法を提案する．データベース，プログラムテーブル，Pivotテーブル等のビジュアルなWebコンポーネントと，Webページの遷移に基づいて実行されるビジネスロジックをドラッグ&ドロップによって結びつけることにより，Webアプリケーションを自動合成する方式について述べる．

Grants-in-Aid for Scientific Research (KAKEN Grants Database @ NII.ac.jp)

Development of a closed-loop subspace identification method with accuracy assurance (Project/Area Number: 21K04124 )
Covariance Analysis of Subspace Identification Methods (Project/Area Number: 15K06146 )
Continuous-time model identification by using adaptive observer (Project/Area Number: 16560385 )
ユニバーサルコントローラの構築 (Project/Area Number: 10750346 )
ユニバーサルコントローラの研究 (Project/Area Number: 08750542 )
自律分散システムの定式化に関する研究 (Project/Area Number: 04218203 )
Search by Researcher Number (80232180)

Associate Professor : Ikeda, Kenji

Research

Personal Web Page

Field of Study

Subject of Study

Book / Paper

Book:

Academic Paper (Judged Full Paper):

Academic Paper (Unrefereed Paper):

Review, Commentary:

Proceeding of International Conference:

Proceeding of Domestic Conference:

Et cetera, Workshop:

Grants-in-Aid for Scientific Research (KAKEN Grants Database @ NII.ac.jp)