Self-Triggered Stochastic MPC With Adaptive Prediction Horizon for Cloud-Based Connected Vehicles Subject to Chance Constraints

This work presents a novel self-triggered stochastic model predictive control (SMPC) scheme for cloud-based vehicle path following control system subject to chance constraints. First, we model the cloud-based vehicle path following control system from a networked stochastic control system perspective. Unlike the conventional periodical sampling approach, a self-triggered mechanism (STM) with adaptive prediction horizon is developed to determine the next sampling time instant and inter-sampling control inputs at each sampling time instant. This mechanism can efficiently reduce the data transmission frequency in the vehicle-cloud communication network, leading to a lower communication load and thus improving the reliability of the system. The STM comprises a set of optimization problems with an adaptive prediction horizon. The optimization problems and threshold design explicitly take the vehicle-cloud communication load into account. Furthermore, a stochastic model predictive control problem with modified constraint tightening and terminal constraint is defined by considering the influence of STM. We develop sufficient conditions to guarantee the closed-loop chance constraints satisfaction in the presence of both adaptive STM and additive disturbances. Then, the recursive feasibility of the optimal control problem and closed-loop stability of the system are investigated. Finally, we illustrate the benefits and effectiveness of the proposed method through numerical examples in vehicle path following control problem.