MPC-based vehicle trajectory planning via distributed ADMM in cloud control systems

In this paper, aiming at the trajectory planning problem of unmanned vehicles in cloud control systems (CCSs), a trajectory planning algorithm based on model predictive control (MPC) is designed. CCSs upload the algorithm to the cloud, then transmit back the planned trajectory, reducing the computational load of the vehicle. The alternating direction method of multipliers (ADMM) algorithm is adopted to accelerate the optimization calculation of the MPC-based vehicle trajectory planning problem. Besides, a distribute ADMM algorithm is proposed by using Cholesky decomposition. The numerical results show that the proposed method enables unmanned vehicles to plan locally optimal trajectories.