Area Coverage of Swarm Robotics Based on Anti-Flocking Framework with Dynamical Clustering

Abstract

Using swarm robots to deal with the task of area coverage has broad applications from regional surveillance, material transportation, to search and rescue. To achieve the goal of higher coverage rate and lower time cost, we present the K-means clustering strategy to execute the dynamical selection of the target position of the next step for each robot, based on the anti-flocking framework. The proposed strategy clusters the non-explored area points to distribute the covering task for each robot and finally assists the swarm to cover the area efficiently. Compared with the classical anti-flocking algorithm, the coverage rate of the proposed method in this paper has been significantly improved. For different combinations of swarm size and interaction radius, this paper uses the genetic algorithm to optimize the model parameters and obtain the optimal model parameters. The simulation results prove that when the model parameters are all optimal, the larger the swarm size is, the more efficient the swarm coverage is, but the interaction radius has no significant effect on the efficiency of the swarm coverage. Finally, we successfully transfer the dynamical clustering algorithm to a semi-physical simulation platform and implement our method adaption in a hardware-in-the-loop simulation to achieve swarm coverage with 7 quadrotors.