Robust ADP-based solution of a class of nonlinear multi-agent systems with input saturation and collision avoidance constraints

Saeed Khankalantary; Iman Izadi; Farid Sheikholeslam

doi:10.1016/j.isatra.2020.07.029

Robust ADP-based solution of a class of nonlinear multi-agent systems with input saturation and collision avoidance constraints

ISA Trans. 2020 Dec:107:52-62. doi: 10.1016/j.isatra.2020.07.029. Epub 2020 Jul 27.

Authors

Saeed Khankalantary¹, Iman Izadi², Farid Sheikholeslam³

Affiliations

¹ Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran. Electronic address: s.khan@ec.iut.ac.ir.
² Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran. Electronic address: iman.izadi@iut.ac.ir.
³ Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran. Electronic address: sheikh@iut.ac.ir.

PMID: 32763017
DOI: 10.1016/j.isatra.2020.07.029

Abstract

In this paper, robust tracking control for a class of nonlinear systems with external disturbance in the presence of input saturation and collision avoidance constraints is investigated. Both of these constraints are considered in a non-quadratic cost function. To obtain optimal control laws, a technique is proposed based on adaptive dynamic programming (ADP) assuming the worst case of disturbance, and without requiring any knowledge of the system dynamics. The proposed algorithm uses neural networks to estimate the optimal cost functions, control policies, and the worst case of disturbance. The Lyapunov theory is used to analyze the convergence of the estimations. Simulation results show the effectiveness of the developed ADP method.

Keywords: Adaptive dynamic programming; Collision avoidance; Input saturation; Multi-agent systems; Optimal control; Robust control.