Cooperative Adaptive Cruise Robust Model Predictive Control Algorithm Research

Adaptive Cruise Control(ACC)is a driving assistance system used for automatically maintaining a safe distance between a vehicle and the vehicle in front on the highway.This aid reduces the driver’s driving burden and improves driving safety.Cooperative Adaptive Cruise Control(CACC),however,is a hybrid of the traditional ACC system in combination with vehicle-to-vehicle(V2V)communication technology.Due to the application of V2 V communication technology,the CACC system exhibits an advantage in reducing the following gap between vehicles and increasing traffic flow throughput.However,communication delay and vehicle dynamics model mismatch can negatively influence the performance of the CACC control system.This thesis presents a CACC control strategy based on Robust Model Predictive Control(RMPC)theory for an intelligent networked fleet that addresses issues of ingestion of dynamics parameter and communication delays.The main research work and innovations of this thesis are as follows:This thesis proposes a CACC control strategy based on the theory of Robust Model Predictive Control(RMPC)for an intelligent networked fleet with dynamics parameter ingestion and communication delays.The control structure of the individual vehicle nodes is divided into an upper level CACC controller and a lower level acceleration controller.The longitudinal dynamics of the vehicle are analyzed,and the lower layer controller compensates for the longitudinal non-linearities of the vehicle by means of the inverse longitudinal dynamics model.Then,the basic idea of robust model prediction and its design framework are presented.A robust model prediction control algorithm for uncertain disturbed systems with input delays is proposed,and its stability is demonstrated.Based on the aforementioned vehicle longitudinal dynamics model,and considering the control objectives of the CACC system,a state-space expression was established for the system with the tracking position error,tracking velocity error,and the acceleration of the ego-vehicle as the system state,and the desired acceleration as the input.The control requirements of the CACC system,such as safety and chord stability,were then analyzed.The control problems that need to be solved by individual vehicle nodes in the CACC formation were formulated as optimization problems.In addition,considering the limitations of vehicle computing resources,a dual-mode control strategy was proposed,which combined offline and online computing.This made it possible to apply the RMPC strategy to the CACC system.Finally,the experimental results from the joint simulation of Carsim/Simulink fully validated that the proposed CACC control strategy in this paper can effectively cope with the impacts of communication delay and dynamic parameter perturbations,ensuring the follow accuracy of the vehicle formation and driving comfort.