Research On Soft-Sensing Technology Of Vehicle Parameters Used For ESP

ESP(Electronic Stability Program) is a new technology of vehicle active security. Itis the evolution of ABS and TCS/ASR. ESP will control the brake system andpower system to improve the vehicle active safety by detecting the driver's drivingintention, condition of road surface and vehicle driving state in an emergency. It isreliable that ESP can provide the guarantee of vehicle active safety in all drivingcondition. Above all, ESP is the latest vehicle technology of avoiding slip in chassiscontrol system.Nowadays, people pay more and more attention to vehicle active safety. ESP hasbecome the key technology of vehicle stability control as the classic of vehiclelateral dynamic control.The vehicle state estimate is one of the significant technologies of ESP. It is boththe research foundation and hotspot of ESP. Moreover, a great deal of researchwork has been done in Europe, Japan and the US, while it is the very beginning inhome.In this paper, some research work of vehicle state estimate has been done on thefoundation of the VOSS(Vehicle Operation Self-organization Simulation)platformdeveloped by Vehicle Operation Simulation Group of Jilin University.I. The acceleration model of motor vehicle is established based on analysisof acceleration character. The Extend Kalman Filters have also been donefor INS and GPS system respectively, and the matrix Q and R aredetermined, too.II. The Data Fusion-Feedback Federal Kalman Filter of GPS/INS system isestablished using Matlab Simulink. The simulation result indicates thatFederal Kalman Filter compensates the accumulated error of INS systemvery well. Furthermore, the accuracy of Federal Kalman Filter proves itsvalidity and necessary.III. The estimate models of vehicle parameters have been done by using theFederal Kalman Filter result and analysing of kinematics and dynamics.a) The vehicle sideslip angle estimate model is established by velocityinformation derived from Federal Kalman Filter and the rotationmatrix from the inertial system co-ordinate to the vehicle chassissystem co-ordinate.b) The definition of tyre sideslip angle model is taken from the analysisof a single-track vehicle model. The chassis and wheels haveidentical velocities at the wheel ground contact points if the tyre slipis small. So the tyre sideslip angles are able to be calculated byforming the velocity balance equations in the vehicle longitudinal andlateral directions.c) The vehicle turning radius model is defined by analysis of centripetalforce and acceleration centripetal.d) The friction co-efficient model has been divided into two differenttypes since the character of friction co-efficient is different betweenthe straight driving and turning states. Under straight drivingcircumstance, the formulas of tyre longitudinal force and dynamicvertical tyre load are used to calculate the friction co-efficient. Whilethe turning state, the Magic Formula is used to determine the tyrelongitudinal and lateral slip. Then the friction co-efficient iscalculated by the corresponding relationship between the tyre slip andfriction co-efficient.IV. The result of enough simulation experiments, such as Lane Change andDouble Lane Change, proves that the design and implement of vehicleparameter estimate models are correct. Under high simulation vehiclespeed, these models perform well during turning and braking. In addition,the long simulation period proves the models' stability and necessary ofFederal Kalman Filter.