Research On Intelligent Vehicle Human-simulated Steering Control Based On MPC Driver Model

With the rapid development of Chinese economy,and the number of cars in the country has also grown rapidly,which makes people’s life more convenient and rich.At the same time,there have been a series of problems such as traffic congestion,frequent traffic accidents and aggravated environmental pollution.However,the intelligent vehicle has effective means to solve problems above.It is the key target of Chinese automobile research and the key link in the research of intelligent transportation systems.The automatic steering of intelligent vehicle is a complex nonlinear,multi-subsystem coordinated control problem involving environmental sensing,decision calculation,executive agency control and other aspects.The steering system is one of the important executive agencies of intelligent vehicle.This paper focuses on the electric power steering system(EPS)of smart cars and its humansimulated control problems include the steering characteristics of skilled drivers,the driver model based on model predictive control(MPC),intelligent vehicle system modeling,novel EPS system design,human-simulated control methods,etc.The main research contents are as follows:Firstly,the paper collected the real-lane steering test data of skilled drivers and the steering test data of driving simulator as the design basis of the human-simulated steering control system.The steering characteristics of skilled drivers are analyzed from the factors affecting the driving trajectory under right-steering and U-turn two typical conditions.The driving trajectory of the skilled drivers in two conditions is fitted based on General Regression Neural Network(GRNN)that to provide foundation for the design of the driver model.Secondly,according to the characteristics that MPC algorithm can predict and optimize the future dynamics of the system at each sampling time,the driving trajectory of the skilled drivers based on GRNN network fitting is used as the reference trajectory to establish the driver model based on MPC considering the skilled drivers steering characteristics.And the driver model based on MPC established in this paper is compared and analyzed with the traditional driver model based on single-point preview from the two aspects of path trajectory and lateral deviation under the two typical conditions.The results show that the driver model based on MPC can track the reference trajectory well under two conditions,which verifies the validity of the driver model.Then,a novel EPS system with double torque/angle sensor is proposed.The novel EPS system dynamic model,motor model,two degrees of freedom model of vehicle,steering resistance torque model,steering wheel torque model and the steering angle feedback control model are established respectively.The human-simulated steering control system is established based on Simulink and CarSim as the simulation platform for the research content of this paper.And the steering test data of actual vehicle and driving simulator of the skilled driver are used as reference trajectories to compare and analyze with the simulation obtained steering wheel angle and steering wheel torque.The results show that the humansimulated steering control system based on Simulink/CarSim proposed in this paper can achieve human-simulated steering better,and verify the effectiveness of the control algorithm designed for the novel EPS system.Finally,based on the linear relationship between the output torque of the magnetic powder brake and the excitation current,the steering resistance torque simulation device is built.The semi-physical platform of the human-simulated steering control system is built based on dSPACE.And the performance tests of the human-simulated steering control system were carried out under two conditions.The experimental results show that the intelligent vehicle human-simulated steering control system proposed in this paper can better realize the human-simulated steering control target,and provides the foundation for the further development and application of intelligent vehicle human-simulated steering control technology.