The present study aims to propose a dynamic interactive self-organizing aggregation (DISA) method for swarm robots. The proposed method was determined by way of the movement of swarm robots, obstacle, and robot sensors. The controller used in the DISA method helps the state selector decide through the utilization of these sensors. Systematic simulations were conducted a different number of robots {10, 25, 50}, different detection radii {3, 4} and different arena sizes {40 × 40, 50 × 50, 60 × 60}. The performance of aggregation behavior was compared with other aggregation methods recommended in literature using Total Distance (TD) between robots, Cluster Metrics (CM), Expected Cluster Size (ECS) metric and aggregation completion time. Moreover, noise at different intensities was applied to sensor inputs of the robots. The robustness of the effect of increasing noise on aggregation behavior was examined comparatively. Consequently, the simulation results based on the other compared methods indicated that the utilization of the proposed DISA method led to a higher performance by 88% in the ECS and CM metric as well as in all TD metric measurements and aggregation completion time results.
Read full abstract