Abstract
〈Vol.2 No.3(2009.5)〉
Titles
▲
■ Design of Free Parameters of State-Dependent Coefficient Form based
on the Relation between State-Dependent Riccati Inequality and Hamilton
Jacobi Inequality
Tokyo Institute of Technology・Yoshihiro Sakayanagi,Shigeki Nakaura,Mitsuji
Sampei
The solvable condition of nonlinear H-infinity control problems is given
by the Hamilton Jacobi inequality (HJI). The state-dependent Riccati inequality
(SDRI) is one of the approaches used to solve the HJI. The SDRI contains
the state-dependent coefficient (SDC) form of a nonlinear system. The SDC
form is not unique. If a poor SDC form is chosen, then there is no solution
for the SDRI. In other words, there exist free parameters of the SDC form
that affect the solution of the SDRI.
This study focuses on the free parameters of the SDC form. First, a representation
of the free parameters of the SDC form is introduced. The SDRI is a sufficient
condition for the HJI, and the free parameters affect the conservativeness
of the SDRI. In addition, a new method for designing the free parameters
that reduces the conservativeness of the SDRI is introduced. Finally, numerical
examples to verify the effect of this method are presented.
▲
■ Model Predictive Sensor Scheduling
Waseda University・Eriko Iwasa and Kenko Uchida
The objective of sensor scheduling is to select a sensor (or a group of
sensors) from multiple sensors at each time step so as to perform optimally
a task based on the sensed data. In this paper, we pose a model predictive
type sensor scheduling problem with a general class of criterion functions
including trace, maximal eigenvalue, and determinant, for discrete-time
linear Gaussian time-varying systems, and develop an approach to solve
this problem based on dynamic programming. We show first that, in a special
case of model predictive deterministic sensor scheduling problem where
the
Riccati recursion of error covariance satisfies a specific structral condition,
the online optimization using the dynamic programming is reduced to a static
optimization, and hence the scheduling algorithm can be easily implemented
online. Next, we pose a model predictive stochastic sensor
scheduling problem to relax the condition of the deterministic sensor scheduling
problem, and show an alternative condition where the optimization is reduced
so that the scheduling algorithm can be easily implemented online. Finally,
we discuss an example to illustrate the two sensor schedulinng algorithms.
▲
■ Estimation of Contact Location from Force Measurements Disturbed by
Noise
Wakayama University・Hiroki Mukai and Kenji Nagase
This paper considers a method to determine a contact point location between
two rigid bodies from a 6-axis force/moment sensor.
In the noiseless case, it is well known that the contact point location
can be determined directly by solving equations representing the force/moment
balances for the exerted force and the resultant sensor signals. Although
we can substitute measured signals to the solution whether noise exists
or not, if noise exists, a minimizing solution to the error of force/moment
balances will be preferable.
In this paper, we formulate the estimation problem as a minimization problem
of the weighted sum of the error of force/moment balances. The optimization
problem is solved analytically and the solution is derived in closed form.
The solution for the noiseless case can be regarded as the minimizing solution
for the error of moment balance by assuming that the force balance holds.The
proposed method is also extended to the case where the measurement signals
for multiple sampling instants can be used. A numerical example and experiments
are shown to prove effectiveness of the proposed method.
▲
■ Shape-Grinding by Direct Position / Force Control
University of Fukui・Guanghua Chen,Nabtesco Corporation・Weiwei Xu and
University of Fukui・Mamoru Ninami
Based on the analysis of the interaction between a manipulator's hand and
a working object, a model representing the constrained dynamics of the
robot is first discussed. The constrained forces are expressed by an algebraic
function of states, input generalized forces, and constraint condition,
and then direct position / force controller without force sensor is proposed
based on the algebraic relation. To give the grinding system the ability
to adapt to any object shape being changed by the grinding, we added a
function estimating the constraint condition in real time for the adaptive
position / force control. Evaluations through simulations by fitting the
changing constraint surface with spline functions, indicate that reliable
position / force control and shape-grinding can be achieved by the proposed
controller.
▲
■ A Frequency Domain Approach of a Robust Control Designby Fractional
Calculus
Nagaoka National College of Technology・Fujio Ikeda and Shigehiro Toyama
This paper deals with a robust design for linear time invariant systems
by fractional order control (FOC), in which controllers described by differential
equations of non-integer orders. The purpose of this paper is to take advantage
of the introduction of the control parameters and satisfy additional specifications
of design, ensuring a robust performance of the controlled system with
respect to gain variations and noise. A method for tuning the controller
is proposed to fulfill five different design specifications. The specifications
about on the gain crossover frequency and the phase margin are readily
satisfied, together with the damping property of the time response of the
controlled system being kept. Simulation results are given to illustrate
the effectiveness of thismethod.
▲
■ Closed Loop Identification Based on the Virtual Reference Feedback Tuning
Applied To a Virtual Two-Degree-of-Freedom Control System
Osaka University・Osamu Kaneko,Yong Kawn,Toshiyuki Ohtsuka
A new identification method with respect to the parameter tuning of a controller is presented.
Here, we introduce a virtual two-degree-of-freedom control structure with a feedforward controller described by using a mathematical model of a plant with a tunable parameter.
After performing a one-shot experiment, we apply the virtual reference
feedback tuning (VRFT) , which is a rational and effective tuning method
for the parameter of a controller with only one-shot experiment data, to
a {\it{virtual}} feedforward controller by using the experimental data
obtained in the actual closed loop.
We give a condition for a prefilter which is applied to the data to guarantee that the obtained parameter using the VRFT of a controller is close to the desired one.
We also show that the prefilter for the identification in the proposed
method has a simpler form than that obtained in the normal VRFT for two-degree-of-freedom
control scheme .
Finally, in order to show the validity of the proposed method, we give an experimental result on the identification of the dynamics of the opening-closing speed of an elevator door.
▲ ■ Delay-Independent Stabilization of Teleoperation Sysetems with Time
Varying Delay
Kanazawa University・Hiroyuki FUJITA and Keio University。Toru NAMERIKAWA
This paper deals with the stability for nonlinear teleoperation systems
with time varying communication delays. The proposed method is the use
of passivity-based controllers with time varying gains which depends on
the rate of change of time varying delay. In our proposed method, stability
condition is independent of the magnitude of the communication delay and
the damping of the robot. The delay-independent stability is shown via
Lyapunov stability methods. Several experimental results show the effectiveness
of our proposed teleoperation structure.
▲
■ Observer-based Identification of Unknown Exogenous Input via Pseudomeasurement
Approach
University of Miyazaki・Takuro Kimura,Akira Ohsumi and Michio Kono
A novel approach to identify the unknown time and magnitude of stepwise or impulsive exogenous input is proposed by introducing an idea of pseudomeasurement with regard to the unknown magnitude.
The efficacy of the proposed approach is verified by simulation studies. This paper forms a deterministic counterpart to ones published elsewhere for stochastic systems.
▲ ■ A New Method for Blind Source Extraction
Kyushu Institute of Technology・Tsubasa YOSHIHARA and Kiyotoshi MATSUOKA
Blind source separation (BSS) is a method for recovering a set of statistically
independent signals from the observation of their mixtures without any
prior knowledge about the mixing process. If, as a special case, only one
source component is to be extracted, it is called blind source extraction
(BSE). Since BSE involves a smaller number of parameters to be estimated
than BSS, it requires less computation time. In this paper we propose a
new algorithm for BSE of the convolutive mixture. The algorithm determines
the extractor by evaluating independence between a target signal and other
signals. It has some good properties. First, since it is formulated in
the time domain, we do not need to worry about the so-called permutation
problem. Second, by applying a particular constraint on the extractor the
signal quality at the sensors is preserved through the extraction process.
A couple of experiments are shown in which the proposed algorithm is applied
to the mixture of five voice signals.
|
