کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

An Accelerated Algorithm for Minimum-Time Multi-Agent Motion Planning

نویسندگان :

الهه نادی¹ ‌Maryam Babazadeh²

1- دانشگاه صنعتی شریف 2- دانشگاه صنعتی شریف

کلمات کلیدی :

Multi-agent motion planning،ADMM،CFS،collision avoidance،distributed optimization

چکیده :

This paper presents a distributed computational framework for minimum-time multi-agent motion planning in complex environments with static obstacles. Traditional motion planning methods face significant computational challenges in multi-agent scenarios, requiring initial guesses close to optimal solutions and suffering from inefficiency at scale. The proposed approach integrates the Alternating Direction Method of Multipliers (ADMM) with the Convex Feasible Set (CFS) algorithm to address these limitations. The framework decomposes the overall problem into independent subproblems for agent dynamics and collision avoidance, facilitating parallel computation and efficient coordination among agents. The CFS algorithm approximates non-convex obstacle constraints as convex sets, transforming computationally challenging problems into tractable quadratic subproblems. Extensive simulations in complex obstacle environments demonstrate that our ADMM-CFS framework achieves remarkable improvements, including 93\% reduction in computation time, superior path quality, and enhanced convergence compared to Interior Point Method (IPM). The method maintains safety constraints while producing smooth, coordinated paths for multiple agents, making it suitable for real-time applications in autonomous systems and multi-robot coordination.