Research On The Railway Track Geometry Surveying Technology Based On Aided INS

The large-scale high-speed railway construction in China brings vast and urgent demands of accurate surveying of the railway track geometry. The existing optical and other surveying methods cannot meet the accuracy and efficiency requirement at the same time, thus finding a fast and accurate surveying method is the only way to solve this problem. The theory and method of railway track survey based on aided inertial navigation system (A-INS) is studied in this thesis, which provides a new solution for rapid and accurate measurement of the railway track geometry.Track geometry measurement is a problem of determining the position and attitude of the rails, focusing on measuring the geometric shape of the track, and is relative measurement in essence. The A-INS has excellent continuous relative measurement capability, and the track geometrical parameters can be measured by A-INS in a non-contact manner through moving on the track. Firstly, the error propagation model of the aided inertial navigation system is established, which provides a theoretical method to analyze the accuracy of the A-INS track measuring trolley.1. A complete error propagation model of the aided inertial navigation system is established for the application of railway track surveying. The A-INS is considered as a stochastic system driven by noise and whose output is also treated as stochastic process. The continuous-time error state Kalman filter of the aided inertial navigation system can be reduced to a linear time-invariant system driven by white noise when reaching steady state performance. Taking Laplace transform of the continuous-time error state Kalman filter functions yields the transfer function of the aided inertial navigation system based on which the relation from the error sources and the surveying error is determined and the time-and spatial-relative accuracy can be studied.2. The stochastic characteristics of the A-INS is studied and measurement accuracy of the track geometrical parameters is quantitatively analyzed based on the established error propagation model of the A-INS. It is proved theoretically that sub-millimeter relative measurement accuracy is achieved by the navigation grade INS with auxiliary information and meets the accuracy demand of the high-speed railway track surveying. The non-holonomic constraint of the track on the A-INS track trolley motion is proved to be a key point for improving the relative measurement accuracy of the A-INS and the simulation results provides a strong proof for this theoretical analysis. The effect of the main A-INS error sources on the railway track geometric parameters measurement accuracy is quantitatively analyzed, which provides a reference for the sensor choice in the A-INS track trolley design. The influence of the moving speed of the trolley and GNSS position update rate on the surveying accuracy is discussed, which is a practical consideration for the real system.The quantitative analysis of the measurement error validates the feasibility of using the A-INS technique to survey railway track geometric parameters of the high-speed line. The second part of the paper focuses on the design and implementation of the railway track geometry surveying system based on A-INS, namely A-INS track trolley, including the hardware design and algorithm development.3. The hardware and software development of the A-INS track trolley are introduced. Key points of hardware design that affecting the A-INS track trolley performance and details on the preprocess of the sensor data are discussed with emphasis. Time synchronization scheme of different sensors and level-arms accurate measurement principle and method are described in detail since data synchronization is the premise of multi-sensor integrated system. The data processing algorithm for the A-INS track surveying trolley is developed. Firstly, The Kalman filter for the aided navigation system is designed, and several practical methods for INS initialization are proposed. Theoretical analysis of A-INS error points out the importance of non-holonomic constraint, thus in the algorithm design, the method and skill of how to use non-holonomic constraint are given. Finally, the method of how to calculate the railway track deviation or deformation and estimate the track geometric state based on the A-INS position measurement is introduced.4. The performance of the A-INS track surveying trolley is comprehensively evaluated and several key points of the theoretical analysis of measurement error are verified through actual railway track surveying experiment.1) The internal accord accuracy of the A-INS track surveying trolley is evaluated by comparing the results of multiple independent surveys on the same track section. On the other hand, the external accord accuracy of the track parameters measurement, such as deformation amplitude, track irregularity of both short and long wavelength, cant and gauge, is evaluated by comparing with the existing high accuracy measuring method.2) The effect of the non-holonomic constraint update on the measurement accuracy of the A-INS track surveying trolley is discussed through actual experiments and is in agreement with the theoretical analysis.3) The surveying performance of the A-INS track surveying trolley on the curved track section is evaluated and shows that the measurement accuracy is the same on different types of track section.4) We discuss the effect of optimal smoothing processing, the results show that optimal smoothing process cannot improve the relative measurement accuracy of A-INS track surveying trolley when operated with reasonable use of non-holonomic constraint update.In summary, this paper focuses on the theoretical and practical engineering problem of using aided inertial navigation system to the railway track geometry surveying, and attempts to provide the theoretical reference and engineering practice guidance for the forthcoming inertial precise surveying.