An Optimised Robot Swarm Flocking with Genetic Algorithm

Collective motion (CM) is a behaviour frequently seen in social animals, such as bird flocks and fish schools. The CM in swarm robotics could be achieved via virtual forces between the robots that coordinate the movements of the swarm. Traditionally, the control parameters of the CM model are fine-tuned empirically to achieve a coordinated movement of robots in a swarm, which is a challenging task and only works for a specific scenario. To address this, the genetic algorithm optimisation method is employed to identify the optimal values for enhancing CM performance and ensuring perfect alignment within the swarm. This paper focuses on two primary objectives to attain superior CM behaviour while conserving energy: i) minimising the virtual forces exerted among the robots and ii) maximising the alignment of the swarm’s headings. The simulation results showed a significant improvement compared to the state-of-the-art CM mechanism. Furthermore, the optimised parameters provide stability to the swarm, enabling it to maintain its structural integrity throughout the simulation.