Preceding Vehicular State Prediction And Multi-Objective Cruise Prediction Control Of Connected Vehicles

Adaptive cruise control(ACC)system is an important part of advanced vehicle assisted driving systems.By controlling throttle opening and braking components,ACC can automatically adjust velocity and effectively reduce the drivers’ burden and the probability of drivers’ misoperation and hence improve driving safety.In ACC systems,the acceleration and deceleration of preceding vehicles is subjective and uncertain.By predicting the acceleration state of preceding vehicles,the host vehicle can do acceleration or deceleration in advance and avoid severe acceleration or deceleration,which improves ride comfort and safety and reduce fuel consumption.Therefore,it is significant to investigate the preceding vehicular state prediction and multi-objective cruise control of connected vehicles.By reviewing and summarizing relevant references at foreign and domestic,the thesis adopts the gaussian learning and model predictive control method to investigate the prediction methods of the preceding vehicular state and multi-objective cruise predictive control of connected vehicles by taking into account such ACC objectives as safety,economy and comfort,The main work of this thesis is listed as follows.(1)Considering the modeling problem of ACC systems,the inter-vehicle kinematics in the ACC system and vehicle dynamics are firstly analysed.Combined with CaiSim software platform,then the longitudinal kinematics model of connected vehicles,vehicle dynamics model,and the vehicle longitudinal dynamics model are established respectively.(2)Considering the multiple conflicting performance indexes and hard prediction of the preceding vehicular acceleration of ACC systems,the Gaussian process regression method is firstly used to learn and predict the preceding vehicular acceleration profiles.Then model predictive control is adopted to achieve the multi-objective control of the ACC system.Therefore,the multi-objective cruise learning predictive controller of connected vehicles is designed based on the preceding vehicular state prediction and is verified by some simulation experiments.(3)Considering the different requirements of following and ride comfort under different driving conditions,the fuzzy control method is used to adjust the weights of the performance indexes of the multi-objective cruise predictive control system.Then the multi-objective cruise predictive controller with varying weights for connected vehicles is designed based on the preceding vehicular state prediction.Some simulation experiments verify the effectiveness of the designed controller.