Abstract: A method of model parameter updating is researched to improve the accuracy of simulation results of high mobility tracked vehicle dynamic model. A dynamic model of high mobility tracked vehicle is established, and the cement road and the gravel road are selected for updating the model parameters according to the statistical regularity of the driving conditions. The simulation results of dynamic model without parameter updating and the corresponding real vehicle test results under the same driving conditions are compared and analyzed, and the expression of objective function for model parameter updating is given. The updating parameters which influence objective function strongly are screened by using orthogonal experiment design method. The radial basis function neural network approximation models about the relation among updating parameters and objective functions are established to solve the issues of large calculation quantity and inefficiency of parameter updating. By analyzing the change rule of objective functions with updating parameters, the dynamic model parameters are updated simultaneously by using the second non-dominated sorting genetic algorithm (NSGA-Ⅱ) for two driving conditions. The final results of parameter updating are obtained. The research results show that the simulation accuracy of high mobility tracked vehicle dynamic model is effectively improved, and the availability of the proposed method is validated.

Key words: ordnance science and technology, high mobility tracked vehicle, parameter updating, multi-objective genetic algorithm, radial basis function neural network