Command Filtered Adaptive Tracking Consensus of Random Nonlinear Multi-Agent Systems

Abstract

This article investigates the topic of adaptive tracking consensus for a family of nonlinear multi-agent systems modeled by random differential equations, which are different from well-known stochastic differential equations. This paper provides some primitive results on random nonlinear multi-agent systems. An improved backstepping method named command filtered control is adopted to derive the adaptive control law, where the convergence of the filtering error is guaranteed. To tackle the serious nonlinearities and uncertainties, a series of dynamic gains are introduced in the design process. The tracking errors for each follower concerning the output of the leader can be regulated arbitrarily to a small enough value by choosing appropriate tuning parameters. All signals of the closed-loop system are analyzed to have bounds almost surely. Moreover, the feasibility of the theory developed in this paper is validated by an example of numerical simulation. Note to Practitioners—Inspired by all kinds of biological clustering or grouping behaviors in our natural world, the research on multi-agent systems has long been popular in both theoretical and engineering scenarios. In addition, colored noises are ubiquitous and negligible when dealing with control problems of different engineering plants. Based on the above observation, this paper was motivated to realize the tracking consensus control of a class of nonlinear multi-agent systems disturbed by colored noise, which is also called random nonlinear systems, with the help of an enhanced adaptive backstepping approach called command filtered control. Although there exist nonlinearities and random noises in each follower agent, the objective of output consensus still could be achieved with the combination of the methods of dynamic gains and command filtering. The application potential of this research is considerable, such as drone formation performance and drone cruise. Our future research will further investigate control problems of random nonlinear multi-agent systems under some practical obstacles such as actuator failures.