Two-Phase Performance Adjustment Approach for Distributed Neuroadaptive Consensus Control of Strict-Feedback Multiagent Systems
针对有向拓扑下严格反馈多智能体系统的实际预设时间无领导者一致性问题，提出一种基于两阶段性能调整的分布式神经自适应控制方案，确保一致性误差在预设时间内收敛到任意小的残差集，且瞬态和稳态性能可调。
This article addresses the practical prescribed-time leaderless consensus problem for multiple networked strict-feedback systems under directed topology. Different from most existing protocols for finite-time consensus that rely on the signum function or fractional power state feedback (thus, the finite convergence time is contingent upon the initial positions of the agents or other design parameters), the proposed distributed neuroadaptive consensus solution is based on a two-phase performance adjustment approach, which exhibits several salient features: 1) the consensus error is ensured to converge to a preassigned arbitrarily small residual set within prescribed time; 2) the tunable transient behavior and desired steady-state control performance of the consensus error is maintained under any unknown initial conditions; and 3) the control scheme involves only one parameter estimation, significantly reducing the design complexity and online computation. Furthermore, we extend the result to practical prescribed-time leader-following consensus control under directed communication topology. Numerical simulation verifies the benefits and efficiency of the proposed method.