| Vehicle active safety is solve by motor vehicle quantity blowout growth and driving technology good and bad are intermingled caused frequent traffic accidents,one of the effective ways to huge casualties and property losses,but the car active safety performance evaluation need to consider the driver behavior characteristics of manipulation,in-vehicle-road under the closed loop system is analyzed.In order to deal with the time-consuming and expensive disadvantages of the real vehicle experiment,the early stage of the development of the active safety system will be carried out on the human-vehicle-road closed loop simulation platform.Therefore,how to establish a driver model with good consistency with the real human driver behavior characteristics is the core problem of the research.This paper is funded by the national key research and development project "research and development and industrialization of intelligent assisted driving technology for electric vehicles".It aims to understand how human drivers perceive the environment and vehicle status information,and make reasonable decisions and actions on this basis.How to learn and master the dynamics characteristics of unfamiliar vehicles was studied.The details are as follows.(1)establish vehicle-road model.The lateral and longitudinal vehicle dynamics models are established.The model of double shift line and S shape road is established by fitting discrete road data points.The vehicle-road model is prepared for the closed-loop simulation verification of the driver model.(2)the driver steering behavior model is established based on the calculation model of brain emotional learning circuit(BEL).The driver steering model drives the BEL model to determine the ideal steering wheel Angle with the far and near visual Angle information of the output of the two-point preview model.The validity of the driver steering model based on BEL is verified by simulation under the double-shift condition.Keeping the parameters of the driver steering model unchanged and changing the longitudinal velocity of the vehicle with a large span,closed-loop simulation verified the good adaptive ability of the BEL model to the longitudinal velocity change of the vehicle.(3)based on the calculation model of brain emotional learning circuit with multi-input and multi-output(MIMO-BEL),the driver model of transverse and longitudinal comprehensive control is established.Referring to the modeling method of transverse driver model,the longitudinal control driver model is established based on BEL model.Finally,a comprehensive driver model based on MIMO-BEL is established by combining transverse and longitudinal driver models,and the simulation results show that the model has good acceleration and path following accuracy.(4)based on Nagendra stable adaptive control,an adaptive driver steering model with the characteristics of human drivers’ adaptive behavior to the change of lateral vehicle model parameters is designed.Human driver in the process of long-term driving,can learn the linear dynamic response characteristics,and to establish the internal reference model,when the vehicle parameters change,the driver under the guidance of internal reference model,through the steering wheel angle and the corresponding dynamic responses of the vehicle,still can be a very good vehicle follow a desired trajectory.In this paper,an adaptive steering model is established based on Narendra stable adaptive control to simulate human drivers’ adaptive ability to vehicle parameter changes.To sum up,based on the research results of psychology and biology and the two-point preview model and the brain emotion learning circuit model,the driver model was established from the cognitive process to the driver model.With reference to the horizontal driver model,the BEL model is driven by longitudinal acceleration following error information to determine the ideal uniform throttle pedal opening.The closed loop simulation experiment verifies that the steering driver model based on BEL has a good adaptive ability to vehicle longitudinal velocity change.The horizontal and vertical driver models are combined to establish the horizontal and vertical integrated control driver model,which eliminates the complex decoupling process of horizontal and vertical control as the horizontal and vertical integrated control driver model based on modern control theory.It is assumed that the driver has enough knowledge of the linear dynamic characteristics of the car through the long-term driving process and forms an internal reference model.Based on Narendra stable adaptive control,an adaptive steering model is established to simulate the human driver’s adaptive ability to vehicle state changes. |
PDF Full Text Request