An Affection-Based Dynamic Leader Selection Model for Formation Control in Multirobot Systems

In this paper, a dynamic leader selection process of a multirobot system with leader-follower strategies is studied in terms of formation control. A fuzzy inference system is employed to evaluate the status of robots by means of their states. Based on the status, an affection-based model is proposed to trigger a leader selection module. Followers send out unsatisfied signals when they are disappointed at the current leader. The abashment value of the leader changes with its own status as well as the number of unsatisfied signals received from its followers. When its abashment value goes beyond a given threshold, a leader reselection process is triggered. Moreover, a swap-greedy algorithm is proposed to approximate the optimal solution for confirming the leader-follower relationship, which can be described as a combinatorial optimization problem to minimize the total travel distance of all the robots. Extensive simulation results demonstrate that the proposed model can improve the probability of a robot team escaping from local extreme points significantly, and even in the case of leader failure, the team can reselect a leader autonomously and keep moving toward the target.