Speed sensor interference source identification based on GWO-CFDP algorithm

In order to accurately identify the sources of interference affecting the TCU speed sensors during train operation, an improved Density Peak Fast Clustering (CFDP) algorithm based on Grey Wolf Optimizer (GWO) was proposed. Initially, the actual interference signals of the train were subjected to feature analysis; following this, feature autoextraction and dimensionality reduction were performed on the processed signals using a two-layer sparse auto encoder network along with kernel principal component analysis. Subsequently, the classification and identification of four types of interference conditions were realized using the proposed GWO-CFDP algorithm. The experimental results demonstrated that the interference source identification algorithm proposed in this paper achieved an identification accuracy of 0.990 and an F1-measure of 0.991 for the four interference conditions, thereby validating the effectiveness and practical value of the algorithm in the field of interference source identification.