Research On The Integration Of CACC Control Strategies In Internet Of Vehicle

With the rise of the Internet of vehicles industry,Intelligent vehicle has attracted a lot of attention.At present,relevant scholars at home and abroad have carried out a lot of research on CACC control strategy,but most of them are only limited to the unilateral control performance of CACC,and few literatures design effective CACC integration control strategy for mixed driving conditions,which can not guarantee the control performance of CACC during the condition switching.It has a great lack of comprehensive consideration of control stability,safety and comfort.In view of the problem,research is conducted from the control method,control strategy design,fusion control strategy and simulation experiment.In order to take into account the overall performance of CACC control,the system and specific theoretical modeling and simulation experiments are carried out around the stable driving condition and dangerous driving condition CACC control strategy,and it is integrated into the full condition CACC control strategy.Under stable driving condition,innovative improvement is made on the control strategy of vehicle longitudinal acceleration and spacing error based on DDPG algorithm controller.Under dangerous driving condition,innovative improvement is made on the control strategy of vehicle acceleration and spacing error and risk time based on network lagrangian system,which is based on support vector machine fusion of the above control strategies.Prescan simulation platform based on driving simulator is built to verify the control strategy.The contents of the paper are as follows:(1)For stable driving conditions,the Priority Value Experience Pool(PVEP)and Sapley Experience Pool(SEP)are added based on the Depth Determination Policy Gradient algorithm(DDPG),and control strategy precise evaluation module,innovative design of CACC hierarchical controller based on Double Experience Pools and Optimization Evaluation Deep Deterministic Policy Gradient(DOE-DDPG).The top layer controls the speed and acceleration of the vehicle by CACC,which directly determines the overall performance of the system,the bottom layer generates the corresponding throttle/pedal opening to control the vehicle.The experimental results show that compared with DDPG,DOE-DDPG control strategy reduces the average acceleration error by 0.12m/s~2,the average spacing error by 0.13m/s~2,and the actual acceleration remains below 0.35m/s~2.(2)Aiming at the dangerous driving conditions,the dynamic model of CACC based on the network lagrangian system is established,and the model graph theory is established to reproduce the actual communication mode of CACC.Combined with the error constraint control,the distributed finite time controller and acceleration and safety distance control strategy are designed innovatively to achieve the convergence of acceleration error and distance error of following vehicle to the state of pilot vehicle and achieve the consistency of CACC coordinated control.The experimental results show that compared with SMPC,the designed distributed finite time control strategy reduces the average acceleration error by 0.11m/s~2,the average spacing error by 0.12m/s~2,and the actual acceleration is maintained below 1.1m/s~2.(3)In view of the mixed driving condition,the framework of CACC fusion control strategy is designed innovatively.According to the input of front vehicle state information,the dominant influence parameters are determined based on the importance weight of objective parameters of Multi-Objective Decision algorithm,and the current driving condition is determined,and the optimal control strategy is matched based on the support vector machine control strategy matcher.When the driving condition changes,the control strategy matching calculation model recalculates the priority of the controller and outputs the optimal control strategy of the current condition in real time.The experimental results show that compared with SMPC,CACC fusion control strategy can reduce risk percentage up to 24.87%,average acceleration error by 0.2m/s~2 and average spacing error by 3.2m.In conclusion,the control performance of CACC is further improved from three working conditions and two levels of stable/dangerous driving mixed driving".The research results enrich the theory and method of intelligent vehicle control.