Identification Of The Key Parameters In Vehicle Dynamics

The aim of the parameters identification of the vehicle dynamics is to find the value of parameters, especially some that can't be acquire directly, neither can be measured directly nor calculated. It is based on the I/O data of experiment and simulation, and use appropriate algorithm. It is significant for reverse engineering and the cost of vehicle development can be reduced.Two problems of this field have been solved in this thesis:First, vehicle inertial parameters identification based on genetic algorithm. The parameter is necessary for vehicle dynamics modeling and simulation. The traditional methods were empirical formula and test bed with shortcomings of low accuracy and high cost. A new approach based on genetic algorithm is promoted here. Vehicle multi-body dynamics model was built based on ADAMS. Special experiments were operated to produce the goal data. The estimation value is calculated with the aim function which is defined as the least difference between system output and goal data. Two methods are used whose results are compared and the performance is analyzed.Second, vehicle side-slip angle identification based on ANN (Artificial Neural Networks). Side-slip angle is the key parameters of the vehicle stability control system. The traditional methods are to get the integral of the acceleration from the non-contract sensor with shortcomings of low accuracy and high cost. A new approach based on ANN was promoted here. The vehicle side-slip angle identification system model was built based on the output data: lateral acceleration and yaw rat. Then testing conditions are estimated and the result that compared with the experiment value is good. At last, the factors that affect the veracity of vehicle side-slip angle identification ANN model are concluded by comparing the estimation result of different model trained by different sample data.Besides, the principle and arithmetic of some ordinary parameters identification technology are concluded, as well as the parameters identification issues of vehicle dynamics and some ordinary vehicle modeling methods.