基于交互式多模型滤波算法机车车速估计

Locomotive speed estimation based on interactive multi-model filtering algorithm

提出了基于交互式多模型滤波算法估计机车运行速度。轮速作为唯一已知的数据作为整个估计系统的输入。由于机车实际运行轨面未知,考虑了三种轨面情况模拟机车轨面运行情况,分别是干燥轨面、潮湿轨面及冰雪轨面。对马尔可夫矩阵进行自适应修正,设计了改进交互式多模型的滤波模型。根据仿真结果表明,不需要额外的轨面识别就能快速精确地估计出机车运行速度。

An interactive multi-model(IMM)filtering algorithm is proposed to estimate locomotive speed.Wheel speed as the only known data input to the entire estimation system.Because the actual running track surface of the locomotive is unknown,three kinds of track surface conditions are considered to simulate the running situation of the locomotive track surface,namely,the dry track surface,the wet track surface and the snow track surface.The adaptive modification of Markov matrix is carried out,and the filter model based on the improved interactive multi-model is designed.According to the experiment result,this method can estimate the locomotive speed quickly and accurately without additional track recognition.