Velocity variation space based multiple agents motion planning for mobile robot

A new approach for multiple agents motion planning is discussed. Different from the previous multi-agent motion planning method, this method is mainly focus on the dynamics constraints and control efficiency. The concept of velocity variation space is introduced to plan the control velocity of agent, and the evaluation function is designed to decide the best control velocity based on the velocity variation space. Simulation experiments show that the proposed method can make agents avoid other obstacles and track the moving target.