Dipole Flow Field for Dependable Path Planning of Multiple Agents

This paper presents a novel approach of path planning for a group of mobile robotic agents working together in a dynamic environment. To do so, the movements of agents are navigated by a vector field of a static flow field and a dynamic dipolar field. The static flow field is formulated based on the anyangle path planning algorithm Theta* with an aim to attract agents to their goals and to avoid obstacles in the map. Meanwhile, the dynamic dipole field pushes agents far away based on their respective moving directions when they tend to collide with other moving agents. The combination of the two fields results in a safe path planning algorithm as well as the deterministic convergence of this algorithm to navigate agents to desired goals. The effectiveness of the proposed approach has been evaluated with simulations.