Composite-Learning-Based Fault-Tolerant Containment Control for Heterogeneous Fully Actuated Multiagent Systems

This article studies the containment control problem for heterogeneous multiagent systems (MASs) with actuator faults and limited information, within the framework of systems engineering for large-scale reliable system design. The problem is decomposed into two sequential processes: distributed estimation via observers and subsequent tracking control for each follower. First, a fully actuated system (FAS) model is derived from a high-order strict-feedback system through variable transformation, and distributed observers are introduced to enable accurate estimation of target states and inputs. Subsequently, a composite-learning-based fault-tolerant control (FTC) method is proposed for each follower, using neural-networks (NNs) to approximate unknown components, including system nonlinearities and fault dynamics. It is shown that all signals in the closed-loop system remain bounded, and the containment error can be reduced to a desired level by properly adjusting design parameters. Numerical simulations on a six-agent system verify the effectiveness of the proposed method and demonstrate its applicability to fault-resilient and practical system deployment.