Path planning and formation control for multi-agent system using ergodic exploration

In this paper, a novel ergodic metric is used to plan paths of multi-agent system in formation based on the ergodic theory. In order to take the full advantages of the formation of multi-agent system in expanding the detection range, obtaining complete environmental information, reducing mutual blockage between agents and performing different types of tasks, we combine leader-follower method and artificial potential field method to control multiple agents and achieve ergodic exploration in formation. To avoid collision between agents and obstacles, a family of vector fields are designed around obstacles so that agents can bypass obstacles. Combining the vector field with the path planning of formation traversal, agents can traverse a given area with obstacles and avoid collide with obstacles.