Multi-Agents Formation Obstacle Avoidance Control Based on Improved Artificial Potential Functions

In multi-agent networks, obstacle avoidance control is a crucial and challenging problem. This paper adopts the Olfati-Saber multi-agent flocking control method to handle obstacle avoidance against spherical obstacles. More specifically, we introduce an improved artificial potential function and virtual target points to model the concerned obstacles. Based on these concepts, an elliptical path is constructed for an agent and an obstacle <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\beta$</tex>-agent located at the ends of the major and minor axes of the ellipse respectively. This ellipse-related approach helps multi-agents to escape local minimum traps and break deadlocks. Moreover, we also integrate model predictive control (MPC) into obstacle avoidance control. This enables the multi-agents to track virtual target points more accurately and efficiently, where MPC can predict the agent states in real-time and optimize their flocking control. Accordingly, the agents can smoothly navigate towards the virtual target points while maintaining a wellcoordinated formation and avoiding collisions. Simulation results show that the improved flocking control combined with MPC is highly effective in driving the agents out of local minima with smooth trajectories and stabilizing the multi-agent system.