基于 Parzen 窗的手动变速器挡位识别方法研究

摘要

During operation a manual transmission vehicle could not obtain the gear position signal without installing the corresponding sensors. This paper analyzed the whole vehicle experiments on the drum tester, read the gear ratio information through CAN bus, then calculated the frequency of each speed ratio signal to determine the number of gear ratios and the range of speed ratio based on the histogram, and the actual speed ratio could be obtained by using the Parzen window function method, which ultimately achieved the accurate identification of the gear position signal. The gear position information acquired by using the Parzen window method agrees well with that obtained directly from the sensor. The comparison verifies the feasibility and accuracy of the method and the paper provides a new method for gear position identification.%针对手动挡车辆在运行过程中无挡位传感器的情况下挡位信号无法识别获取的问题，通过对多次整车转鼓试验进行分析，发现读取 CAN 总线里面的速比信息，然后基于直方图计算各速比范围信号出现的频率，并以此确定挡位数目和速比的大小范围，再利用 Parzen 窗函数的方法获取各挡位的实际速比，可以实现对挡位信号精准的识别。试验结果表明，采用 Parzen 窗的挡位识别方法能够识别出挡位信号。对比 Parzen 窗获取的挡位识别信息与挡位信号传感器直接获取的挡位信息，发现两者的相似度很高，这也验证了该方法的可行性和准确性。该方法提供了整车运行过程中一种挡位识别的新方式。