A neuroendocrine inspired dynamic leader selection model in formation control for multi-robot system

Abstract

Formation control is widely used in industrial, military and daily applications. The development of multi-robot system (MRS) provides new tools and theories to solve this problem. Among related works, the leader-follower strategy is famous for its simplicity and efficiency. Based on leader graphic theory, this paper proposes a neuroendocrine system inspired dynamic leader selection model (NeDLSM) to autonomously switch leaders as such helping the team get out of the woods. A fuzzy C-means algorithm is employed to simulate a nervous system in NeDLSM to evaluate states of robots by fusing external information and status of robots themselves. The neuroendocrine regulation mechanism is referred to adjust two kinds of hormones, which are defined as the specific properties of the leader and the followers, respectively. A superior concentration of hormones will trigger the leader re-selection process. Numerical simulation results show that the MRS with NeDLSM can decrease the convergence error significantly and maintain a required formation in a complex environment.