LAUSR

This letter studies the minimum time formation control issue of autonomous underwater vehicles (AUVs), subject to switching topology during the formation procedure. We first employ the smoothstep function to describe the smooth transition of communication topology. Then, a minimum time formation control problem is constructed by minimizing the integral temporal difference errors of AUVs. To solve the above problem, an integral reinforcement learning (IRL) based controller is designed for AUVs to achieve leader-follower formation task. Of note, the smooth transition in this letter can reduce oscillation and improve safety during the formation change procedure, and meanwhile, the IRL-based formation controller can achieve minimum time formation task in an unknown environment. Finally, the theory solutions are verified by simulation and experimental results.