Faculty Profiles - Tsubakino, Daisuke

Presentations -

Division display 61 - 80 of about 85 ／ All the affair displays >>

A control law for a quadrotor to land on a running ground vehicle under input constraints

Yuki Fukuda, Daisuke Tsubakino

ある低ランク層間構造を保存する階層化最適制御問題に関する一考察

椿野大輔，原辰次

第6回計測自動制御学会制御部門マルチシンポジウム (MSCS2019)

Single control law for tracking a ground vehicle with a quadrotor

Yuki Fukuda, Daisuke Tsubakino

Design of a Control Law for Separated Flows over a Circular Cylinder based on Model Predictive Control and Linear Regression

Yasuo Sasaki, Daisuke Tsubakino

Boundary Observer Design for Sloshing in Fuel Tanks Based on an Approximate Model

行列代数に基づく階層化最適制御の理論とその応用 Invited

椿野大輔

制御理論合宿2018

On Feedback Controller Design for Fluid Flows Based on Approximation

Daisuke Tsubakino, Yasuo Sasaki, Naruya Ito

Quadrotor formation flying based on hierarchical optimal control strategies

Yusuke Kamitani, Daisuke Tsubakino

62nd Annual Conference of the Institute of Systems, Control and Information Engineers

Model Predictive Control of a Separated Flow Around a Circular Cylinder at a Low Reynolds Number

Yasuo Sasaki, Daisuke Tsubakino

SICE Annual Conference 2017

Nonlinear Hovering Control Laws for Small Vectored-Thrust Tail-Sitter UAVs International conference

Kouzi Miyazaki, Daisuke Tsubakino

1st IEEE Conference on Control Technology and Applications

Predictor-feedback for multi-input LTI systems with distinct delays International conference

Daisuke Tsubakino, Tiago Roux Oliveira, Miroslav Krstic

2015 American Control Conference

Event date： 2015.7

Language：English Presentation type：Oral presentation (general)

Country：United States

In this paper, we propose a predictor-based state feedback controller for multi-input linear time-invariant systems with different time delays in each individual input channel. The controller is designed based on the backstepping method. Since the conventional backstepping transformation is not applicable to the systems due to the differences among delays, a modified transformation is introduced. This transformation enables us to design an exponentially stabilizing controller under which the plant behaves as if the delay were absent after a finite time interval. In addition, an explicit Lyapunov functional is constructed. The performance of the controller is demonstrated by a numerical simulation.

Multiparameter extremum seeking with output delays International conference

Tiago Roux Oliveira, Miroslav Krstic, Daisuke Tsubakino

2015 American Control Conference

Event date： 2015.7

Language：English Presentation type：Oral presentation (general)

Country：United States

In this paper, we generalize the previous results of multivariable extremum seeking by assuming output delays not considered before in the control system. The phase compensation of the dither signals and the inclusion of an averaging-based predictor in the closed-loop feedback allow to obtain exponential convergence results to a small neighborhood of the optimal point, even in the presence of delays. The stability analysis is carried out using backstepping transformation and averaging in infinite dimensions. The proposed approach also removes the dependence of the convergence rate on the unknown Hessian of the nonlinear map to be optimized, being user-assignable as in Newton-based extremum seeking results. The control performance is evaluated by numerical simulations.

State estimation for boundary-coupled parabolic distributed parameter systems using spatial averages International conference

Daisuke Tsubakino, Yuh Yamashita

10th Asian Control Conference

Event date： 2015.6

Language：English Presentation type：Oral presentation (general)

Country：Malaysia

This paper proposes an indirect method for designing a Luenberger-type state observer for systems described by boundary-coupled two parabolic partial differential equations. In particular, we concentrate on the case where outputs are spatial averages of the state variables. The proposed method consists of two steps. We first introduce a state transformation that converts the system into another system with boundary measurement. Then, the backstepping approach can provide a state observer for the converted system. An estimate of the original state variables is constructed through the inverse transformation. The performance of the proposed observer is confirmed by a numerical simulation.

Backstepping-forwarding control of parabolic PDEs with partially separable kernels International conference

Daisuke Tsubakino, Federico Bribiesca Argomedo, Miroslav Krstic

53rd IEEE Conference on Decision and Control

Event date： 2014.12

Language：English Presentation type：Oral presentation (general)

Country：United States

This paper considers a boundary control problem of parabolic partial differential equations (PDEs) with an integral of the state variable over the whole spatial domain. To handle this non-strict feedback term, we use a Fredholm-type state transformation that involves two integral terms with kernels defined on the lower and the upper triangular domains. It is shown that the transformation converts the original system with an associated state feedback law into an exponentially stable target system, if the kernels are solutions to a boundary value problem of coupled second-order hyperbolic PDEs. Moreover, the upper kernel is separable if a coefficient of the original system is characterized by a certain parameterized ordinary integro-differential equation. Our approach also provides an explicit feedback gain in a special case. The effectiveness of the proposed controller is confirmed by numerical simulation.

Collision-free formation control of two-wheeled vehicles based on triangular LQ optimal regulator International conference

Takashi Kusunoki, Daisuke Tsubakino, Yuh Yamashita

2014 IEEE Conference on Control Applications

Event date： 2014.9

Language：English Presentation type：Oral presentation (general)

Country：France

This paper considers a formation control problem of two-wheeled vehicles. We first design a base controller based on the structured linear quadratic optimal regulator. The resulting optimal control gain is a lower triangular matrix; thus, agents communicate with each other in a top-down manner. However, vehicles may collide with each other under this control law since it is linear. To avoid collision of vehicles, we design a potential function by deforming the quadratic form of the stabilizing solution of the Riccati equation. The gradient of the potential function gives a control law that guarantees collision avoidance. The control law coincides with the optimal control law if vehicles are far from each other. The effectiveness of the proposed controller is confirmed by numerical simulation.

Boundary control of a cascade of two parabolic PDEs with different diffusion coefficients International conference

Daisuke Tsubakino, Miroslav Krstic, Yuh Yamashita

52nd IEEE Conference on Decision and Control

Event date： 2013.12

Language：English Presentation type：Oral presentation (general)

Country：Italy

In this paper, a state feedback control law is proposed for boundary coupled two different one-dimensional parabolic partial differential equations (PDEs). The control input is assumed to act only at a boundary of one side. The feedback law is designed based on the backstepping method. As a novelty of this paper, a state transformation involving a `discontinuous' integral kernel is introduced in order to handle the difference of diffusion coefficients. This enables us to develop a well-posed backstepping framework for cascades of parabolic PDEs and an exponentially stabilizing controller is derived. The performance of the proposed controller is also validated by numerical simulation.

An algebraic approach to hierarchical LQR synthesis for large-scale dynamical systems International conference

Daisuke Tsubakino, Taiki Yoshioka, Shinji Hara

9th Asian Control Conference

Event date： 2013.6

Language：English Presentation type：Oral presentation (general)

Country：Turkey

This paper considers a linear quadratic optimal hierarchical control problem for large-scale dynamical systems modeled by an interconnected system under multi-scale information exchange networks. We first propose an algebraic characterization of hierarchies by using semigroups the Kronecker product. The multiplication rule of the Kronecker product quite fits to the property of semigroups. As a result, a condition under which the stabilizing solution of the Riccati equation inherits the hierarchy is obtained with the aid of the previous result. Furthermore, the proposed framework makes it possible to understand several previous results on decentralized optimal control from a unified viewpoint.

Backstepping control for parabolic PDEs with in-domain actuation International conference

Daisuke Tsubakino, Miroslav Krstic, Shinji Hara

2012 American Control Conference

Event date： 2012.6

Language：English Presentation type：Oral presentation (general)

Country：Canada

The backstepping method is a systematic design tool for boundary control of various types of partial differential equations (PDEs). There has been no attempt to apply it to PDEs whose input is not at the boundary. In this paper, we consider a problem of feedback stabilization of 1-dimensional parabolic (unstable) PDEs with internal actuation based on the backstepping method. Since such a PDE can not be converted to a stable PDE by state feedback and the state transformation used in backstepping, an additional transformation is introduced. Under a certain condition, the newly proposed transformation moves the input from the interior of the domain to the boundary. This enables us to cancel the residual term that causes the open-loop instability by using the input. Furthermore, this transformation is continuously invertible. Therefore, a stabilizing state feedback for the original PDE is derived through the inverse transformation. The results are demonstrated by a numerical simulation.

Backstepping Observer Using Weighted Spatial Average for 1-Dimensional Parabolic Distributed Parameter Systems International conference

Daisuke Tsubakino, Shinji Hara

18th IFAC World Congress

Event date： 2011.8

Language：English Presentation type：Oral presentation (general)

Country：Italy

This paper is concerned with designing a backstepping-based observer for a class of 1-dimensional parabolic distributed parameter systems whose output is a weighted spatial average of the state. We first show that an integral transformation converts the original system into another parabolic system with boundary observation if the weighting function satisfies a parameterized ordinary differential equation. Then the backstepping observer for the transformed system is available and an estimate of the original state is obtained through the inverse transformation. We also show that the estimation error exponentially converges to 0 in terms of the L2 norm with arbitrary decay rate. Furthermore, a closed-form expression of the observer is provided for systems described by linear reaction-diffusion equations.

Eigenvector-based characterization for hierarchical multi-agent dynamical systems with low rank interconnection International conference

Daisuke Tsubakino, Shinji Hara

2010 IEEE International Conference on Control Applications

Event date： 2010.9

Language：English Presentation type：Oral presentation (general)

Country：Japan

In this paper, we investigate hierarchical multi-agent dynamical systems with low rank interconnection based on a new concept of eigen-interconnection. The eigenvector plays an important role in our framework. We first define the eigen-interconnection matrix for the rank one case. Then, we show eigenvalue distribution of a hierarchical interconnection with rank one eigen-interconnection matrix. The obtained eigenvalue distribution has explicit parameter dependency and the result allows us to place the eigenvalue selectively. Thus it gives useful knowledge to design hierarchical interconnection structure. This concept is naturally extended to the rank two case and a similar result for eigenvalue distribution is obtained. Those results completely include the previous researches as special cases. Furthermore, our characterization has a potential of straightforward extension to the higher rank cases. Finally, we show a numerical example where a hierarchical interconnection structure is designed based on our result.

　 PREV - NEXT 　

<Tsubakino, Daisuke>

Presentations -