• International Journal of Control, Automation, and Systems
• /
• v.1 no.2
• /
• pp.163-170
• /
• 2003

The optimal Bayesian filter for a single target is known to provide the best tracking performance in a cluttered environment. However, its main drawback is the increase in memory size and computation quantity over time. In this paper, the inevitable predicament of the optimal Bayesian filter is resolved in a suboptimal fashion through the use of a receding horizon strategy. As a result, the problems of memory and computational requirements are diminished. As a priori information, the horizon initial state is estimated from the validated measurements on the receding horizon. Consequently, the suboptimal algorithm proposed allows for real time implementation.

• International Journal of Control, Automation, and Systems
• /
• v.2 no.4
• /
• pp.475-484
• /
• 2004

In this paper, we present a dynamic output-feedback receding horizon $H_{\infty}$controller for linear discrete-time systems with disturbance. The controller is obtained numerically from the finite horizon output-feedback $H_{\infty}$optimization problem, which is, in fact, hardly solved analytically. Under a matrix inequality condition on the terminal weighting matrix, the monotonic decreasing property of the cost is shown. This property guarantees both the closed-loop stability and the $H_{\infty}$norm bound. Then, we extend the proposed design method to a reference tracking problem and a problem for time-varying systems. Numerical examples are given to illustrate the performance of the proposed controller.

• Transactions on Control, Automation and Systems Engineering
• /
• v.3 no.3
• /
• pp.159-163
• /
• 2001

An input constrained receding horizon predictive control algorithm for uncertain systems with nonzero set points is proposed. for constant nonzero set points, models with uncertainty can be converted into an augmented incremental system through the use of integrators and the problem is transformed into a zero-state regulation problem for the incremental system. But the original constraints on inputs are converted into constraints on the sum of control inputs at each time instants, which have not been dealt in earlier constrained robust receding horizon control problems. Recursive state bounding technique and worst case minimizing strategy developed in earlier works are applied to the augmented incremental system to yield an offset error free controller. The resulting algorithm is formulated so that it can be solved using LP.

• 제어로봇시스템학회:학술대회논문집
• /
• 1995.10a
• /
• pp.424-427
• /
• 1995

In this paper, a control law based on the receding horizon concept which robustly stabilizes time-varying discrete linear systems, is proposed. A dynamic game problem minimizing the worst case performance, is adopted as an optimization problem which should be resolved at every current time. The objective of the proposed control law is to guarantee the closed loop stability and the infinite horizon $H^{\infty}$ norm bound. It is shown that the objective can be achieved by selecting the proper terminal weighting matrices which satisfy the inequality conditions proposed in this paper. An example is included to illustrate the results..

• Journal of Institute of Control, Robotics and Systems
• /
• v.11 no.9
• /
• pp.755-760
• /
• 2005

This paper proposes an improved implementation algorithm for the continuous-time receding horizon control (RHC). The proposed algorithm has a feature that it has better control performance than the existing algorithm. Main idea of the proposed algorithm is that we can approximate the original RHC problem better by assuming the predicted input trajectory on the prediction horizon has a continuous form, which is constructed from linear interpolation of finite number of vectors. This, in turn, leads to improved control performance. We derive a predictor such that it takes linear interpolation into account and proposes the method by which we can express the cost exactly. Through simulation study fur an inverted pendulum, we illustrate that the proposed algorithm has the better control performance than the existing one.

• 제어로봇시스템학회:학술대회논문집
• /
• 1992.10b
• /
• pp.49-55
• /
• 1992

In this paper, we developed a Receding Horizon Predictive Control for Stochastic state space models(RHPCS). RHPCS was designed to minimize a quadratic cost function. RHPCS consists of Receding Horizon Tracking Control(RHTC) and a state observer. It was shown that RHPCS is equivalent to Generalized Predictive Control(GPC) when the underlying state space model is equivalent to the I/O model used in the design of GPC. The equivalence between GPC and RHPCS was shown through. the comparison of the transfer functions of the two controllers. RHPCS provides a time-invarient optimal control law for systems for which GPC can not be used. The stability properties of RHPCS was derived. From the GPC's equivalence to RHPCS, the stability properties of GPC were shown to be the same as those for RHTC.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.159.1-159
• /
• 2001

A new time-domain FIR parameter estimation called the receding horizon least square estimation (RHLSE) is suggested for stochastic systems by combining the well known least square estimation with the receding horizon strategy. It can be always obtained without the requirement of any \textit{a priori} information about the horizon initial parameter. A fast algorithm for the suggested estimation is also presented which is remarkable in the view of computational advantage and simple implementation. It is shown that the proposed estimation is robust against temporary modeling uncertainties due to their FIR structure through simulation studies.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.444-447
• /
• 2004

A new approach to compute the control moves in Extended Dynamic Matrix Control (EDMC) is presented. In this approach, the number of variables, determined in the inner loop of the control algorithm using iterative methods, is reduced from P , the prediction horizon to M , the control horizon. Since M is usually much smaller than P , this modifies the control algorithm from computational point of view. To justify the modification, the computational requirements are compared to those of the existing EDMC algorithm.

• 제어로봇시스템학회:학술대회논문집
• /
• 1987.10a
• /
• pp.734-740
• /
• 1987

Various kinds of predictive control design methods such as MAC(Model Algorithmic Control), DMC(Dynamic Matrix Control), MC(Extended Horizon Adaptive Control), GPC(Generalized Predictive Control), RHTC(Receding Horizon Tracking Controller), and PVC(PreView Controller) are surveyed and compared in this paper. In addition, stability properties of these control laws known to date are summarized and some new stability results are presented.

• Journal of applied mathematics & informatics
• /
• v.25 no.1_2
• /
• pp.231-244
• /
• 2007

In this paper, we use measure theory for considering asymptotically stable of an autonomous system [1] of first order nonlinear ordinary differential equations(ODE's). First, we define a nonlinear infinite-horizon optimal control problem related to the ODE. Then, by a suitable change of variable, we transform the problem to a finite-horizon nonlinear optimal control problem. Then, the problem is modified into one consisting of the minimization of a linear functional over a set of positive Radon measures. The optimal measure is approximated by a finite combination of atomic measures and the problem converted to a finite-dimensional linear programming problem. The solution to this linear programming problem is used to find a piecewise-constant control, and by using the approximated control signals, we obtain the approximate trajectories and the error functional related to it. Finally the approximated trajectories and error functional is used to for considering asymptotically stable of the original problem.