基于S-method时频分析的制动意图识别方法

Brake Intent Recognition Method based on S-method Time Frequency Analysis

为了提高电动汽车制动能量回收效率,不可避免需要增加制动意图判断的准确率,提出了基于S-method时频分析的制动意图识别方法。鉴于S-method处理非平稳信号兼顾算法响应时间和时频分辨率的优点,通过建立S-method时频模型的基础上,运用奇异值分解方法提取特征值。最后,运用聚类算法识别驾驶员制动意图。实验结果表明,该方法能实时有效识别中等制动和平缓制动。

In order to improve the braking energy recovery efficiency of electric vehicles, it is necessary to increase the accuracy of braking intent judgment, a method of braking intent recognition based on S-method time frequency analysis is proposed. S-method handles non-stable signals to take into account the algorithm response time and time frequency resolution, the characteristic values are extracted by using the singular value decomposition method on the basis of establishing the S-method time frequency model. Finally, the clustering algorithm is used to identify the braking intent. The results show that the method can effectively identify the medium brake and the slow braking.