Cooperative Control of Multi-Agent Systems

This project proposes a simple, efficient and scalable method for motion planning and control of multi-agent systems (MAS) with collision avoidance capabilities. This report begins with the review of studies and algorithms in the fields of cooperative control of MAS and swarm robotics for collective behavior. In this project, classic Artificial Potential Field (APF) algorithm is utilized to implement control and motion planning for autonomous robots (agents) with obstacle avoidance capabilities. The APF algorithm provides effective and smooth trajectory planning compared with other algorithms that are studied in this project. Furthermore, to address classical APF algorithm limitations such as local minima and narrow channels, a Lyapunov theorem approach, is used to optimize potential fields and resolve those limitations. In addition, a special case of modified APF algorithm is studied to implement leader-follower scenario. The results for each controller are verified using MATLAB simulation. Finally, a survey of existing hardware platform to implement MAS formation control algorithms is presented.