Research On Neural-network Adaptive Sliding Mode Control Of Longitudinal Dynamic Behavior Of Intelligent Vehicle

Intelligent vehicle is a high-tech carrier which integrates environmental perception,planning and decision-making,motion control and other functions,it can realize autonomous driving function in different road environment,it also represents an important direction for the future development of automobile.Vehicle dynamic behavior control is one of the critical steps for us to realize autonomous driving of intelligent vehicle.Also,longitudinal dynamic behavior control is of great significance for real-time adjustment of vehicle longitudinal motion state and guarantee of intelligent vehicle driving safety.Intelligent vehicle is a kind of nonholonomic motion constraint system with high nonlinearity and parameter uncertainty,the design of its longitudinal dynamics control strategy is facing higher challenges.In order to further improve the accuracy of longitudinal motion control of smart cars,this paper uses RBF neural network adaptive sliding mode control strategy to study the longitudinal dynamic control of intelligent vehicles.The main research contents of this paper are as follows:Firstly,the longitudinal dynamic model of intelligent vehicle is constructed.Based on the existing test data,the longitudinal dynamic model of intelligent vehicle including engine,transmission and whole vehicle is established systematically,the operation law of throttle actuator and brake actuator can be mastered.On this basis,the inverse model of vehicle longitudinal dynamics is established.At the same time,the reasonable switching logic of throttle and brake is established,which lays a model foundation for the follow-up design of intelligent vehicle longitudinal dynamics control.Secondly,a non-singular terminal sliding mode control strategy is designed.On the basis of fully mastering the principle of sliding mode control,the optimal design of the system control algorithm can be completed by setting the appropriate system input and output.Then based on MATLAB / Simulink platform,the simulation model of longitudinal dynamics sliding mode control of intelligent vehicle is built.The simulation results show that compared with PID control strategy,the non-singular terminal sliding mode control strategy reduces the average tracking error of longitudinal speed more than 50%,it effectively improves the longitudinal speed control accuracy of intelligent vehicle.Thirdly,the longitudinal dynamics control strategy of intelligent vehicle based on RBF neural network adaptive sliding mode is studied.According to the chattering phenomenon of sliding mode control and the approximation characteristic of RBF neural network,the switching gain of non-singular terminal sliding mode control is adaptively adjusted by neural network.The simulation results show that the RBF neural network algorithm can effectively improve the chattering phenomenon of traditional sliding mode control,and significantly improve the longitudinal motion control quality of intelligent vehicles.Finally,based on the D2 P platform,the real vehicle test of the longitudinal dynamics control performance of the intelligent vehicle is completed.In the test,the RBF neural network adaptive sliding mode control algorithm is written into the D2 P rapid control prototype system.Combined with the test vehicle,the real vehicle test of intelligent vehicle longitudinal dynamics control is completed.The real vehicle test results show that the designed neural network adaptive sliding mode control strategy has better control accuracy in the actual application process.It verifies the effectiveness of the control strategy proposed in this paper,and provides rich experience and reference for the design and application of intelligent vehicle longitudinal dynamics control.