Adaptive safety critical compensation control for nonlinear jump systems with its application.

Abstract

In this article, the problem of adaptive safety critical tracking control for nonlinear Markovian jump systems (NMJSs) is addressed, and an event-observer-based compensation strategy is proposed. The main innovation of the presented approach is to effectively cope with the unexpected actuator faults, lumped disturbances, nonlinear uncertainties, and limited computation resources for the NMJSs. In the most existing compensation schemes, only the bias fault without jump dynamics is addressed. Here, the combined faults, disturbances and limited measurable conditions are considered. To solve the limited variable measurable problem, an integrated dynamic event-based state observer scheme is constructed. With the help of state estimation error, a high-order lumped disturbance observation scheme is established without the prior knowledge of disturbance. By virtue of observation values, an event-observer-based system performance recovery controller is proposed. In this control framework, to achieve the disturbance rejection and tracking convergence purposes, an observer-aided sliding mode surface is established. What's more, to reduce the communication burden, a novel dynamic trigger scheme in control channel and a multi-choice-based trigger condition in sensor channel are proposed, and the controller is reconstructed with the event choice. Finally, two numerical examples and the application to a robot system show the safety compensation abilities of the developed approach.