Predefined-time adaptive fuzzy echo state network containment control of uncertain multiagent systems with prescribed performance

Abstract

Most control techniques that put constraining conditions to tracking performance usually involve the aid of barrier Lyapunov functions or integral-type functions to regulate tracking errors within a region surrounded by constants and coordinate axes, which may result in algebraic loop or singular problems. Aiming to drive the entire individuals of the multiagent systems (MASs) into a convex hull composed of multiple leaders with a preset performance freely modulated by decision-maker, in this paper, a predefined-time adaptive fuzzy echo state network containment control agreement with prescribed performance is developed for uncertain MASs subject to input saturation. A fuzzy echo state network, as a syncretism and escalation of conventional radial neural network, is utilized to approximate unknown nonlinear dynamics. A new log-type function is defined via combining coordinate transformation with the trait of the funnel function. Through applying the predefined-time Lyapunov stability criterion, theoretical analysis indicates that all signals of closed-loop network MASs are semiglobally practically predefined time bounded, and the errors evolve within the prescribed boundary customized by a funnel function in a predetermined time. Finally, the practicality of the presented approach is validated through an actual simulation example.