Research On Vehicle Active Collision Avoidance Control Strategy Based On State Estimation Method

With the improvement of vehicle intelligence,vehicle active safety technology has become more and more perfect,and people have gradually increased their acceptance and expectations for vehicle intelligence.As the core technology of vehicle advanced driver assistance system,vehicle active collision avoidance mainly depends on the accuracy of perception system and the assembly rate of vehicle assistance system.Under the low assembly rate of automotive advanced auxiliary systems,the perception system obtains less information,and the state estimation algorithm can be used to obtain vehicle state information,which indirectly improves the perception accuracy and control accuracy of the vehicle.In this paper,a vehicle active collision avoidance system is designed,the vehicle state and road adhesion coefficient are estimated,the vehicle braking strategy on different roads is considered,the horizontal and vertical warning based on V2 X technology is applied,and the horizontal and vertical joint avoidance based on driver comfort is applied.The collision research was carried out,and the joint simulation experiment was carried out.The main research contents are as follows:(1)The overall framework of the active collision avoidance system based on the estimation method of vehicle state and road adhesion coefficient is established.According to the function and principle of the active collision avoidance system,the method of obtaining the vehicle status information in the collision avoidance system is determined,and the vehicle active collision avoidance warning is realized by considering the horizontal and vertical directions.The lower layer obtains the vehicle’s throttle valve,brake pressure and steering wheel angle to realize the vehicle’s active collision avoidance control.(2)The vehicle state estimation and road adhesion coefficient estimation modules in the active collision avoidance algorithm are established.The tire model and the vehicle four-degree-of-freedom dynamic model are the basic models of the estimation module.By simulating and comparing the mechanical properties of the three tire models,the Pacejka tire model,the improved Pacejka tire model and the Dugoff tire model,the most suitable tire model is selected.Based on the dual-volume Kalman filter combined with the tire model and the vehicle dynamics model to estimate the vehicle state and road adhesion coefficient,the estimated precision error of the average adhesion coefficient of the four wheels on the fixed road and the variable road is less than 9%.The accuracy error is 6%smaller than the extended Kalman filter.(3)Two modules of early warning system and control strategy in the active collision avoidance system are built.In the early warning system,the blind spot of traditional sensors is considered and V2 X technology is used to solve this problem to realize the full state observation of the environmental vehicle.According to the V2 X communication of the vehicle,and the fuzzy control algorithm is used to obtain the steering wheel angle of the environmental vehicle and the azimuth angle of the vehicle to judge the danger of the vehicle.In the control decision module,the control strategy of active collision avoidance of vehicles on different roads is considered,the road adhesion coefficient is integrated into the safety distance model,and the thresholds of vehicle warning and braking are set on the basis of considering the driving characteristics of the driver and extended to It is convenient for graded braking early warning decision-making,and the acceleration changes caused by graded braking are optimized by applying cubic polynomial.Through the simulation analysis of different road surfaces,lane-changing distances and vehicle speeds,the minimum lane-changing distance for the vehicle to maintain vehicle stability is obtained.Combined with the horizontal and vertical warning and control decision-making of the vehicle,the horizontal and vertical fusion is used to avoid collision,and the vehicle is slightly braked during the steering process to reduce the vehicle’s lane-changing distance and ensure driving safety and comfort.(4)Design the lateral and longitudinal controllers of the vehicle’s active collision avoidance control system.The longitudinal controller uses the PID algorithm to take the expected acceleration of the vehicle as input,calculates the expected throttle opening of the vehicle through the vehicle inverse dynamics model and the throttle opening characteristic curve,and realizes the longitudinal control of the vehicle;the lateral controller uses the model prediction theory,Based on the establishment of a two-degree-of-freedom vehicle model,the prediction of the future state is realized by rolling optimization,and the steering wheel angle of the vehicle is calculated as the input of the lateral control of the vehicle.(5)Build a co-simulation platform of vehicle active collision avoidance system based on Prescan,Carsim and Simulink,and conduct co-simulation experiments.CCRs,CCRm and CCRb scenarios and front car cut-in scenarios are established to verify the effectiveness of the active collision avoidance control strategy,and the simulation results show that the active collision avoidance algorithm in this paper can warn according to the dangerous state of the vehicle,and make reasonable lane change or braking operation.Vehicle lateral following error is within 15%,the vehicle longitudinal braking process uses multi-stage brake deceleration,respectively,3m/s2 and 7.5m/s2,the system in the premise of ensuring the safety of the occupants,to a certain extent to ensure the comfort of the occupants.