An Information-Driven Algorithm in Flocking Systems for an Improved Obstacle Avoidance

The research field of swarm robotics continues to draw inspiration from the behavior of animals in nature. Obstacle avoidance and the navigation in unknown areas are major problems in control of swarming agents. In recent years, a large number of algorithms have been developed for this purpose. These algorithms are mainly based on artificial potential functions and many of the existing strategies do not enable agents to escape from non-convex obstacles. Due to the issue of local minimum, agents with a limited sensor range can often stay stuck behind concave obstacles. In this study, we propose a new algorithm to make many concave obstacles avoidable for flocks in unknown environments through sharing and evaluating the aggregated sensing information among the group.