Research On Longitudinal Control Strategy Of Vehicles Considering Driving Style And Slope

Traditional vehicle longitudinal control methods usually only consider the static characteristics of the road and the vehicle,without fully considering the driver’s personalized driving style and real-time changes in road conditions.As a result,they are unable to achieve longitudinal control that adapts to driving styles and changes in slope.To address this issue,this study aims to develop a new longitudinal control model that combines driving style characteristics,vehicle driving slope and vehicle control to achieve driving style-aware and slope adaptive vehicle longitudinal control.First of all,in the feature selection stage of driver style recognition,this paper preprocesses the data using boxplot method and adopts the whale optimization algorithm based on the sigmoid function to carry out feature selection for the natural driving dataset,providing a data basis for driver style recognition.In the driver style recognition stage,a spectral clustering method based on Bi-LSTM autoencoder is used for driver style classification and recognition.The experimental results of feature selection and driver style recognition on natural driving data show that the proposed method can achieve a feature recognition accuracy of 97.34% on a significantly reduced number of features,and the selected driving data features can reflect the driving behavior characteristics of drivers.Sample analysis of feature data shows that the proposed method can use selected features to effectively reflect differences in driving styles among drivers.Secondly,based on the vehicle dynamics model,this study builds the speed controller for the longitudinal control model of the vehicle,by building the drive/brake system and engine model,for further analysis of the impact of driving style and grade on the longitudinal control of the vehicle.Carsim combined with MATLAB/Simulink are used to construct the vehicle speed ramp signal and sine signal to verify the effectiveness of the constructed model,showing that the speed controller has high control accuracy with speed control errors of 0.11m/s and 0.36m/s,respectively.Afterwards,to incorporate the driving style recognition results and road slope into the vehicle control,an incremental model predictive control model is employed,with the adjustment coefficient expressing the impact of different driving styles on the longitudinal control of the vehicle added to the constraint conditions.The fuzzy logic is also used to consider the weight of the control quantity in the objective function with respect to driving style,road slope,jerk,and vehicle speed.Objective function and constraint conditions are transformed into a quadratic programming problem to solve the desired acceleration and complete the construction of the longitudinal control model of the vehicle.Finally,an adaptive model predictive control strategy simulation platform considering driving style and slope is built in this study,and the model is preliminarily validated in virtual working conditions of flat ground and 5%road slope.The preliminary validation results showed that the proposed longitudinal control model of the vehicle can reach and stabilize near the reference speed set within a short time.The acceleration of the vehicle in the process of reaching the stable speed is within a reasonable range,which ensures the comfort of the driving process,and the vehicle’s tracking of the reference speed is good enough to reflect the operational state of the vehicle.To consider the influence of real driving conditions,a slope calculation method based on GPS altitude and vehicle speed is used,and the recognition results of the corresponding slope and driving style are input into the actual working conditions constructed using actual driving data.Compared with traditional PID control,the simulation results show that the proposed model combined with driving style and road slope has an average tracking error of 0.03147 m/s for vehicle speed,lower than the PID control error,and can still maintain good stability and tracking accuracy in typical slope sections.