基于遗传算法的汽车ESP液压系统参数辨识

Parameter Identification of Automotive ESP Hydraulic System Based on Genetic Algorithm

ESP液压系统是保证汽车行驶安全性的重要执行机构,根据ESP液压系统的结构原理,在Matlab/Simulink环境下建立其实时仿真模型。采用实车测试环境下记录的试验数据作为参考数据,基于遗传算法对ESP液压系统模型的未知特征参数进行辨识。对比运行于实时仿真机的ESP液压系统模型输出的制动轮缸压力与实车测试环境下的记录值,结果表明用辨识数据参数化的ESP液压系统模型具有较高的符合度,可满足构建ESP系统硬件在环仿真测试环境的需求。

Abstract ： ESP hydraulic system is an important actuator for the active safety control system of vehicle. In the Matlab/Simulink environment, the real-time simulation model of ESP hydraulic system, which contains unknown characteristic parameters, is established on the strength of its structural principle. In order to identify the unknown characteristic parameters, several test cases that cover passive and active pressurization modes of the model are executed in the real vehicle experiment environment, and the collected vehicle measurements are divided into two groups. The unknown characteristic parameters of the model are identified with genetic algorithm by using one group as reference data. The model is compiled and downloaded to the real-time simulation system, which is named dSPACE AutoBox. Using the valve currents and pump motor signals of another group of measurements as the inputs of the model, the brake wheel cylinder pressures of the same group of measurements are compared with the outputs of the model. The results show that the real-time simulation model of ESP hydraulic system with the identification results of characteristic parameters has high accuracy, and can meet the requirements of establishing hardware in the loop simulation test environment for ESP system.