Research On Combined Measurement Method Based On Kalamn Filtering And Its Application In Railway Inspection

Track Irregularities directly affect the safety of vehicle running and the passengers’ riding comfort. In order to detect irregularities, most countries use special track inspection car for dynamic measurement of the railway geometry parameters. The key of railway inspection is to establish a measurement benchmark with high accuracy.In comparison to different benchmark establishing methods at home and abroad and with the knowledge of the inertial measurement technology for the aerospace and other military fields, the thesis mainly discusses a kind of combined measurement technique based on Kalman filtering, which can be used to establish measurement benchmark of the railway geometry parameters.Based on the analysis of the characteristics of the line, the thesis models its vertical and horizontal characteristics of a typical trajectory, and output signals of the corresponding accelerometers and angular rate gyros and gives out simulations.The key of an inertial measurement system is the solution of the attitude matrix. The thesis uses the quaternions and the rotation vector approach to solve the attitude matrix. The quaternion approach is a mathematical method using four components to represent a space vector. The attitude matrix can be used to represent the mutual rotate relationship between two vectors. The Picard attitude matrix approximation algorithm processes angular increment, when the gyro’s output is the angle increments, it can be used directly, but the gyro’s output of the system is the angular rates, so the rotation vector method is used to extract the angular increments. After calculating the attitude matrix and according to the elements of the attitude matrix, then the attitude angles can be calculated. Furthermore, the forces of accelerometers measuring in the carrier coordinate need to be converted into the geographic coordinate in order to calculate the vehicle speed of east, north and vertical direction according to the multiplication principle, finally the track inspection car’s running trajectory can be gotten by numerical integration solution.Inertial measurement system is of high accuracy at the beginning of work, but with time increasing, the solver error will be getting larger and directly impacts on the accuracy of measurement benchmark. In order to compensate for measurement errors, the thesis designs a combined measurement system based on Kalman filtering technique. Using the speed and location error values betweent GPS and inertial measurement system as external variables, and establishing state equations of a combination of the measurement system with GPS and inertial measurement error models, an optimal estimation of the inertial measurement error can be executed with Kalman algorithm, and then the outputs of the inertial measurement system can be compensated by way of an open-loop correction.To further verify the validity and correctness of the combined measurement system, the thesis establishes an experimental system based on virtual instruments, and conducts a field experiment. Through the analysis and comparison of the measured data, the results show that the designed combined measurement system is feasible and effective in establishing benchmark for measuring track parameters. In a conclusion, the designed combined measurement can significantly improve the accuracy of inertial measurement system.