Driving track clustering method based on railway signals and Beidou positioning

The invention discloses a driving track clustering method based on railway signals and Beidou positioning. The method comprises the following steps: reading position and speed information related to atrain set in railway control and interlocking information transmitted by a computer interlocking system, predicting the train set position at the next moment in real time by adopting a Kalman filtering algorithm, and comparing the train set position with Beidou positioning information of the train set transmitted by vehicle-mounted equipment; removing the Beidou positioning information as noise if the Beidou positioning information deviates from the predicted position of the train by more than a specified distance at the same moment; otherwise, storing the position point corresponding to thecorresponding Beidou positioning information; based on the mode, in the section driving process, accumulating a certain number of position points, then optimizing the position pointsthrough a binary k-means algorithm to obtain