A competition-based scheme using k-WTA networks for multi-target pursuit and encirclement in dynamic environments

The pursuit–evasion problem in multi-robot systems (MRS) has raised much attention in the last two decades. However, most of them are focused on single-target pursuit problem and the cooperative mechanism is adopted. To fill this gap, in this paper, we propose a competition-based scheme for multi-target pursuit and encirclement in dynamic environments. To assign task more efficiently, a multi-target k-winners-take-all algorithm (MK-WTA) is proposed. For obstacle avoidance in original particle-based model, an obstacle-aware particle-based model (OAPM) is developed. Combining MK-WTA with OAPM, a complete framework is formed for solving multi-target pursuit problem. To testify the performance of the proposed algorithm, extensive simulations are conducted as well as the comparative studies with other existing methods. The effectiveness of the proposed scheme in the real world is validated by physical experiments with differential-driven-wheeled robots.