Emergent Behavior Detection and Task Coordination for Multiagent Systems

Four aspects of the technical solutions in the book are presented: First, the impact of interaction dynamics on the convergence conditions related to interaction topologies is discussed, utilizing a discontinuous cooperative control algorithm of updated design. Second, distributed least-squares and Kalman filtering algorithms for agents with limited interactions are elaborated upon. Third, a general framework of distributed nonlinear control is established, and distributed adaptive control for nonlinear systems with more general uncertainties is presented. Based on the proposed framework, a distributed nonlinear controller is designed to deal with task coordination of robotic systems with nonholonomic constraints. Finally, the problem of optimal multiagent task coordination is addressed and solutions based on approximate dynamic programming and approximate distributed gradient estimation are presented.