摘要:

Sliding mode control (SMC) of a general class of nonlinear control systems is considered in this work. The conventional SMC technique requires knowledge of the upperbounds of disturbances and modelling uncertainties to assure robustness. However, this may not be easy to obtain. As a remedy, an estimation process for these dynamic perturbations is employed jointly with the SMC technique. This new methodology, sliding mode control with perturbation estimation (SMCPE), offers a robust feedback control with much lower gains than its conventional counterparts against slowly varying perturbations. This resolves one of the problematic issues which has caused concern over the years of development of SMC applications. An interesting perspective of selecting the cut-off frequency for s dynamics is presented with a novel upperbound argument. Much desirable tracking fidelity is arrived through SMCPE in the computer simulation studies for a two-link manipulator. A companion approach to SMCPE, the discrete equivalent of it, is also studied in this paper. A discrete SMC controller is developed for the same class of systems and perturbations using the discrete attractivity condition. This method is shown to possess properties similar to the continuous SMCPE.

ISSN:0020-7179    

DOI:10.1080/00207179208934350

出版商:Taylor & Francis Group    

年代:1992

数据来源: Taylor

摘要:

We present observer-based controller designs for decentralized systems that are reliable to sets of control-channel outages. The designs guarantee stability and aH∞norm bound on the closed-loop system despite complete sensor or actuator outages for any subset of a preselected set of control channels. The results are presented in a unified discrete/continuous setting. We show how the continuous-time and discrete-time results can be generated from the unified formulation and present an algorithm for solving the discrete-time design equations by making use of the delta-operator formulation.

ISSN:0020-7179    

DOI:10.1080/00207179208934351

出版商:Taylor & Francis Group    

年代:1992

数据来源: Taylor

摘要:

In this paper, the decentralized minimal controller synthesis (DMCS) algorithm is shown to be extremely effective in controlling nonlinear multivariable plant with unknown parameters. Both the stability and robustness of this new adaptive algorithm are investigated. It is shown that if the plant coupling terms and nonlinearities are slowly varying, compared with each adaptive loop bandwidth, then the DMCS algorithm yields an asymptotically stable closed-loop system. However, if the coupling terms and/or nonlinearities are rapidly varying, tracking errors are shown to be bounded. An estimate of the upper bound of error convergence is derived. Implementation studies of the algorithm applied to a two-degree-of-freedom manipulator are included. The results of these experiments are closely aligned with the theoretical predictions.

ISSN:0020-7179    

DOI:10.1080/00207179208934352

出版商:Taylor & Francis Group    

年代:1992

数据来源: Taylor

摘要:

The differential matrix Riccati equation with periodic coefficients is considered. It is shown by means of a counter-example that a sufficient condition for the existence of an infinite number of periodic non-negative definite solutions given in a previous paper is not appropriately stated. A suitable adjustment of the assumptions allows one to work out rigorous sufficient conditions. Thanks to these results, a necessary and sufficient condition for the existence of the periodic strong solution, which relied on the incorrect condition, can be shown to be still valid.

ISSN:0020-7179    

DOI:10.1080/00207179208934353

出版商:Taylor & Francis Group    

年代:1992

数据来源: Taylor