Research On Multi-Source Information Fusion Techniques For Direct Measurement Of Track Distortion And Horizontal Alignment

With the continuous expansion of China’s railway transportation network and the increasing train operating speeds,improving the accuracy of railway track geometry parameter measurement has become increasingly critical.Currently,China still heavily relies on a single high-precision inclinometer sensor for measuring track horizontal parameters.However,this sensor is highly dependent on the inclinometer sensors manufactured by JEWELL company in the United States,and its electromagnetic damping technology only reduces the effects of impact oscillations,while issues such as dynamic time phase lag and abnormal large values still persist.Therefore,this paper proposes an innovative measurement solution based on multi-sensor information fusion.By adding an angle sensor to work in conjunction with the inclinometer sensor,the track horizontal parameters can be calibrated to improve measurement accuracy.Firstly,a multi-sensor synchronous measurement scheme for track horizontal was constructed.It was found that the track inspection device can increase the direct measurement of track torsion parameters by angle sensors.In order to ensure the dynamic performance of the measurement,the measurement characteristics of the sensor were analyzed.By introducing the track torsion contrast difference algorithm,the track inversion level was obtained,laying the foundation for the solution of the existing dynamic measurement problem of track horizontal through direct measurement and inversion level multi-source information fusion.Secondly,a thorough analysis and research of data waveform segmentation was conducted.Taking guessed torsion and measured torsion as research objects,a waveform segmentation scheme strategy and process were formulated,and the DTW algorithm was improved to make it suitable for data processing of track inspection device horizontal parameters.The results of the guessed torsion segmentation were mapped to the measured horizontal level to achieve the goal of waveform segmentation of the measured horizontal level.Finally,a track horizontal multi-sensor information fusion algorithm was formed.For different segments of measured horizontal values and their corresponding inversion levels,an information fusion method was studied to eliminate time-delay defects and abnormal values contained in the measured horizontal level.Experimental tests were conducted to verify the applicability and effectiveness of multi-sensor information fusion technology.The application of the research results in this paper can effectively alleviate the dependence of the track inspection device on the high-precision tilt sensor from JEWELL company in the United States and significantly improve the dynamic performance of track horizontal measurement.