Bio-inspired adaptive PSO for swarm robotics: balancing search efficiency and physical limitations for dynamic target scenarios

Searching for multiple moving targets in unknown environments with limited resources is a challenging problem for a robot swarm. Many existing methods face difficulties in simultaneously optimizing search accuracy, energy efficiency, and adaptability. To address this, this paper introduces an enhanced adaptive multi-swarm particle swarm optimization (AMSPSO) framework. This work introduces three key technical contributions: a revised particle update mechanism incorporating real-world communication and energy constraints; a dynamic sub-swarm division mechanism (DSDM) that allocates robots to targets based on real-time search performance; and an energy-aware inertia weight adjustment strategy based on the Student’s t-distribution to balance exploration and exploitation. In comprehensive simulations across eight complex scenarios, AMSPSO achieved an average performance improvement of approximately 41% in energy efficiency-oriented search efficiency and a reduction of about 20% in energy consumption relative to the next-best performing algorithm in our tests. These results suggest that AMSPSO is a promising approach for the problem of dynamic multi-target search under communication and energy constraints.