Physical Validation of PSO and ACO Based Algorithms for Multi-Agent Robotic Systems

This work focuses on the physical validation of Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) based algorithms for trajectory planning and goal searching tasks with multi-agent robotic systems. We used Pololu 3Pi+ differential robots with added wireless communication capabilities, and we ran our experiments on a 3.8 × 4.8 m platform with 6 OptiTrack motion capture cameras. Our methods consider the kinematics of differential robots and use PID based controllers to ensure smooth trajectories and safe motor speeds for the robots. In our PSO experiments, the swarms were able to reach goal positions under different conditions. In our ACO experiments, we used virtual agents to find a suitable trajectory to the goal in the test space, and a physical robot was able to follow the calculated trajectory.