Research On Multi-objective Adaptive Cruise Control Algorithm Considering Road Conditions

At present,the development of intelligent driving is rapid,and the Advanced Assisted Driving System（ADAS）,as an important part of intelligent driving,is more and more accepted by people.Among them,the Adaptive Cruise Control（ACC）as an advanced assisted driving system developed earlier and is more mature.In part,it has had an important impact on driving safety and driving comfort.However,the current ACC system activation conditions have strict requirements on the road environment,and require more intervention by the driver to continuously adjust the inter-vehicle time gap.In order to further expand the application scenarios of the ACC system and liberate some of the driver’s manipulation,this paper studies a multi-objective ACC system that adapts to the full-line,high,medium,and low adhesion coefficient roads,which further improves driving safety,comfort and economy.Aiming at the real-time estimation of road adhesion coefficient,this paper establishes a seven-degree-of-freedom vehicle model（7-DOF）and a Dugoff tire model,which are used as the state observation basis for the road adhesion coefficient estimator.Two nonlinear estimators are established: Extended Kalman Filter（EKF）and Unscented Kalman Filter（UKF）;and the accuracy of the two estimators were verified in Carsim software.Finally,the unscented Kalman filter with higher accuracy is selected to estimate the road adhesion coefficient in real time.Aiming at the establishment of the variable vehicle-to-vehicle time distance model,the UAV is used to take aerial video of road traffic flow in different road adhesion conditions,and the stable car-following behavior is extracted from the Tracker software through video playback,and 360 groups of stable car-following behaviors are obtained by removing the mutation abnormal data using the 3σ principles.Based on the carfollowing data,a polynomial fitting method is used to establish a model of the relationship between the variable headway,the road adhesion coefficient and the speed of the host vehicle.Aiming at ACC control,a controller with upper and lower layers is established.The upper planning layer is divided into a longitudinal acceleration planning layer and a lateral additional yaw moment planning layer.Model predictive control is used longitudinally according to spacing deviation,speed deviation,acceleration deviation and jerk deviation.The longitudinal desired acceleration with multiple objectives is planned,and two sliding-mode controllers are established laterally according to the yaw rate deviation and the center of mass sideslip deviation.At the same time,a fusion model based on vehicle stability is established to calculate the lateral total additional yaw moment;the lower controller Acceleration conversion and torque distribution are carried out,and the multi-objective longitudinal acceleration planned by the upper-layer controller is converted into the expected total torque of the whole vehicle by using the vehicle longitudinal dynamics model,and PID feedback control is added to compensate the longitudinal modeling error.The optimal torque distribution model is established to convert the total longitudinal torque and lateral additional yaw moment into the expected torque of the four wheels of the vehicle.At the same time,in order to ensure that the four wheels avoid slip,the slip rate PI control is added to obtain the adjustment of the four wheels.The torque,the final output torque of the four wheels is obtained by combining the adjustment torque with the distributed expected torque,which is input to the in-wheel motor or braking system.Aiming at the multi-objective nature of ACC,using the characteristics of iterative optimization of genetic algorithm,the genetic algorithm is combined with the weight coefficient in model predictive control to solve the weight distribution coefficient of the optimal cost function under different working conditions,and the fuzzy switching rule is used to adjust the motion state of the two vehicles in real time.The weight coefficient achieves multi-objective control,which ultimately ensures that the vehicle has good following,comfort and economy under the premise of horizontal and vertical safety.In this paper,Carsim and Matlab/Simulink platforms are used to verify the effectiveness of the algorithm,and dynamic scenarios of different linear roads,roads with different adhesion coefficients and different working conditions are established in Carsim.Compared with the PID controller and the fixed weight MPC controller respectively,the results show that the algorithm has better safety,road adaptability and working condition adaptability.