An Optimized Flocking Motion with Attention Module for Obstacle Avoidance

A flocking control algorithm with attention module for obstacle avoidance is proposed, which is based on the classical rules of Reynolds. To enable the self-organized swarm having better environmental adaptability, the coefficients of social force are determined as variables, optimized by using the non-dominated sorting genetic algorithm II (NSGA-II). In this work, two different obstacle avoidance models are elaborated, one is with obstacle attention module and the other is with potential field module. Then, the experiment for the self-organized swarm with tasks moving towards the target zone without obstacle collisions is set up and the two models are compared in this scene. From the results, the flocking control model with attention module shows better performance in collisions and motion consistency, nearly 32.35% improvement in time cost and a 30.94% improvement in aggregation degree.