Controller-Dynamic-Linearization-Based Distributed Model-Free Adaptive Control for Nonlinear Multiagent Systems

The leaderless or leader-following consensus tracking, along with containment control problems of nonlinear multiagent systems (MASs) using controller-dynamic-linearization-based distributed model-free adaptive control (CDL-DMFAC) method are addressed in this article. By virtue of dynamic linearization (DL) technology, the distributed output and ideal controller of MASs are first converted to the corresponding equivalent DL data models, respectively. Then, a pure data-based CDL-DMFAC scheme is uniformly constructed by employing I/O data regardless of the state space model of MASs. The convergence analysis is rigorously proved by a designed data energy function without using global topology graph information. Furthermore, the control strategy and convergence result are extended to acrlong MIMO MASs. Finally, extensive simulations are performed to verify the validity of theoretical results.

I. Introduction

For complex distributed systems that each intelligent agent exchanges information via a band-limited digital network and cooperates to accomplish collaborative tasks, multiagent systems (MASs) have been involved in numerous fields, including robotics [1], [2], unmanned aerial vehicles [3], etc. In contrast to traditional point-to-point systems, MASs have the superiorities of low cost, high-fault tolerance, potential intelligence, as well as strong reliability in various cooperative control tasks, such as consensus tracking [4], containment control [5], formation control [6], rendezvous [7], flocking [8], and so on. Under leaderless consensus tracking framework, the states of all agents are required to be consistent finally. If the outputs of all agents can track the leader’s trajectory consistently, it is called leader-following consensus. For containment control, which could be recognized as the multiple leaders case of the leader-following consensus, all followers should be included in a convex hull [9], [10], [11]. The other control tasks, such as formation control, could be easily tackled with above mentioned three collaborative objectives. Consequently, how to realize containment and consensus control of MASs are fundamental research topic.