Researches On The Control Of Lateral Assistance For Intelligent Vehicle And Hardware-in-the-loop

As a new round of technological revolution and the trend of industrial transformation,automotive intelligence has become a strategic direction for the development of the global automotive industry,and it also provides a variety of solutions to the transportation problems facing the industry.Vehicle collision is one of the most serious traffic problems at present.Collision Avoidance Assistance,one of the intelligent technologies of automobiles,assists the driver to change the state of the vehicle to avoid the collision or mitigate the harm caused by it.The AEB system has been recognized by various parties and has played a significant role in longitudinal collision avoidance.However,in emergency conditions with short obstacle avoidance distances,high vehicle speeds,and low overlap rates,lateral collision avoidance can avoid more accidents.Therefore,from the perspective of improving driving safety,research on lateral collision avoidance assistance control strategy is more important.This paper focuses on the research of intelligent vehicle lateral collision avoidance assistance control strategy.It analyzes three aspects of collision risk assessment and collision avoidance trajectory planning,driver steering torque modeling,and lateral collision avoidance assistance control strategy.Lateral collision avoidance assistance control strategy hardwarein-the-loop test platform,based on specific scenarios,perform driver change test to test and verify the algorithm.(1)Design of collision risk assessment and collision avoidance trajectory planning.Based on the characteristics of the host vehicle motion and traffic vehicles motion under the risk scenario,this paper combines the trajectory prediction method based on the kinematics model and the trajectory prediction direction based on the driver’s intention to design the final predicted trajectory.From the perspective of possible collision,the main vehicle’s own lane and adjacent lanes are evaluated,and the degree of danger of the current state of the host vehicle is determined based on the TTC and the vertical and horizontal spacing of the vehicle.When the vehicle is in danger of collision,the collision avoidance trajectory is derived based on the different structured road environment of the host vehicle,based on the polynomial of the fifth degree of one yuan.The collision avoidance trajectory is limited to a certain range based on the three conditions of allowed lateral acceleration,target orbit environment,and obstacle width.(2)Modeling of driver’s steering torque in emergency situations.In order to obtain the driver’s steering torque model which can be used in MPC controller,this paper models the driver’s steering torque during emergency steering.By analyzing the action process of driver’s steering torque,it can be divided into driving arm torque,feedforward torque and feedback torque.Firstly,a driver’s arm model is established based on a single-degree-of-freedom inertial-damping-stiffness mechanical system,and a driver torque-steering wheel angle acquisition system is established.Design three different collection conditions to analyze the influencing factors of the driver’s arm system parameters,and identify the mechanical characteristics of the arm model.Secondly,for the driver’s feedforward torque,a linear driver’s feedforward steering torque model is established based on the idea of preview.A CarSimbased driving simulator was built to collect driver data and establish different elementary danger scenarios.The relevant dangerous factors in the case of a car were analyzed,and the dangerous working conditions were established based on this.Collect the driver data to identify the model parameters,so as to obtain the identification result of the corresponding driver.Finally,the feedback torque modeling is divided into tire feedback torque and steering system feedback torque.(3)Lateral Collision Avoidance Assistance Control Algorithm.The algorithm used in this paper is model predictive control including predictive model,rolling optimization and feedback correction.The closed-loop steering model for human-vehicle-road construction includes body dynamics model,tire dynamics model,steering system and driver model,and combined with Taylor expansion to approximate linearization.Based on the above model,a lateral collision avoidance assistance controller based on the model predictive control algorithm is designed.An objective function is designed that comprehensively considers the accuracy of the tracking trajectory and the stable operation of the controlled object under the control amount.The design takes into account the constraints of actuator performance and vehicle stability.Finally,the above controller is modeled in Simulink and a driver model based on MPC is established.In the vehicle simulation model software,simulation is performed under typical working conditions for algorithm debugging and verification,and validity of controller is initially verified.(4)System integration and hardware-in-the-loop experiment analysis.Based on the dSPACE real-time test tool chain,a hardware-in-the-loop test platform for lateral collision avoidance assistance control algorithm is established.Use ASM,MotionDesk and Matlab software to build suitable scenarios and match vehicle models.Use Simulator,iBooster and SensoWheel and other hardware to build a driving simulation environment.Integrate collision risk assessment,trajectory planning,and collision avoidance assistance controllers into the platform.A variety of typical dangerous working conditions were set up and real-time driverin-the-loop experiments were conducted on this basis to test and verify the overall effect of the control strategy.The results show that the lateral assistance obstacle avoidance module can accurately complete the risk assessment and trajectory planning,and effectively assist the driver in lateral collision avoidance.