Search

CN-122015865-A - Rail transit vehicle contactless positioning system based on electromagnetic induction

CN122015865ACN 122015865 ACN122015865 ACN 122015865ACN-122015865-A

Abstract

The invention provides a non-contact positioning system for rail transit vehicles based on electromagnetic induction, which comprises a distributed magnetostrictive electromagnetic induction array module, a dynamic magnetic bias compensation module, a double-phase difference calculation positioning module, a time-space domain calibration feedback module and a vehicle-mounted trackside double-end verification module. The invention adopts a five-layer modularized architecture, all modules work cooperatively to form a full-flow processing link from original signal acquisition to final position output, and positioning precision and stability are ensured through multi-algorithm fusion calculation and closed-loop design, so that sub-meter-level non-contact continuous positioning of rail transit vehicles is realized.

Inventors

  • SUN JINMIN

Assignees

  • 雄安京雄快轨有限公司

Dates

Publication Date
20260512
Application Date
20260310

Claims (10)

  1. 1. A rail transit vehicle contactless positioning system based on electromagnetic induction, comprising: the distributed magnetostrictive electromagnetic induction array module is used for synchronously collecting magnetostrictive deformation electric signals and electromagnetic induction voltage signals to form a double-path original signal set; The dynamic magnetic bias compensation module is used for receiving the double-path original signal set, identifying and counteracting electromagnetic interference in the double-path original signal set and generating a calibrated double-signal set; The double phase difference resolving and positioning module is internally provided with a complex frequency domain phase unwrapping algorithm and an array weighted pseudo-range projection algorithm, and is used for receiving the calibrated double signal set, and carrying out phase unwrapping and position resolving on the calibrated double signal set through bidirectional fusion and mutual feedback of the two algorithms to generate an initial three-dimensional position coordinate of the train and a double phase difference resolving error value; The time-space domain calibration feedback module is internally provided with a time-space domain sparse Bayesian optimization algorithm and a multi-source heterogeneous data fusion self-adaptive variable order Kalman filtering algorithm, and is used for acquiring the GIS space-time data and the GIS real-time working condition data of the track traffic network, receiving the initial three-dimensional position coordinates of the train and the double-phase difference calculation error value, performing time-space domain calibration on the initial three-dimensional position coordinates of the train through bidirectional fusion and mutual feedback of the two algorithms, and generating calibrated three-dimensional position coordinates and position correction coefficients of the train; The vehicle-mounted trackside double-end verification module is used for receiving the three-dimensional position coordinates and the position correction coefficient of the calibrated train, acquiring the induction signals acquired by the trackside fixed stake unit, calculating to obtain trackside real-time position data, carrying out weighted fusion verification on the trackside real-time position data and the three-dimensional position coordinates of the calibrated train to generate a final positioning result of the train, and simultaneously calculating an error feedback value and feeding the error feedback value back to the dynamic magnetic bias compensation module and the double-phase difference calculation positioning module respectively.
  2. 2. The electromagnetic induction-based rail transit vehicle contactless positioning system according to claim 1, wherein in the double-phase difference solution positioning module, a complex frequency domain phase unwrapping algorithm converts a calibrated double-signal set from a time domain to a complex frequency domain, extracts double-phase differences and generates an unwrapped double-phase difference matrix through filtering unwrapping, an array weighted pseudo-range projection algorithm converts phase differences in the unwrapped double-phase difference matrix into array unit pseudo-range values, weights the array unit pseudo-range values by matching corresponding space weight coefficients, projects the weighted array unit pseudo-range values to a unified geographic coordinate system, and performs weighted fusion to generate initial three-dimensional position coordinates of a train, and meanwhile calculates pseudo-range residual reverse feedback to the complex frequency domain phase unwrapping algorithm.
  3. 3. The electromagnetic induction-based rail transit vehicle contactless positioning system is characterized in that a bidirectional fusion and mutual feedback process of two algorithms in a double-phase difference resolving and positioning module is that an array weighted pseudo-range projection algorithm is based on a double-phase difference matrix after unwrapping output by a complex frequency domain phase unwrapping algorithm, a pseudo-range residual feedback value is calculated and obtained, the pseudo-range residual feedback value is reversely transmitted to the complex frequency domain phase unwrapping algorithm and used for adjusting complex frequency domain filter coefficients of the pseudo-range residual feedback value and eliminating phase unwrapping accumulated errors, and the double-phase difference matrix after unwrapping optimized by the complex frequency domain phase unwrapping algorithm is input into the array weighted pseudo-range projection algorithm again to iteratively optimize resolving accuracy of initial three-dimensional position coordinates of a train.
  4. 4. The electromagnetic induction-based track traffic vehicle contactless positioning system according to claim 1 is characterized in that in a time-space domain calibration feedback module, a time-space domain sparse Bayesian optimization algorithm is used for carrying out L1 norm sparsification on track traffic network GIS space-time data and track real-time working condition data after splicing, bayesian optimization is carried out by combining a double-phase difference calculation error value, an optimized space-time data matrix is generated, a noise covariance initial value is calculated, a self-adaptive variable order Kalman filtering algorithm for multi-source heterogeneous data fusion is used for taking the noise covariance initial value as a process noise covariance initial value, a filtering order is self-adaptively adjusted according to train speed, a filtering equation is constructed to complete prediction and update, a calibrated train three-dimensional position coordinate is generated by calibrating an initial three-dimensional position coordinate, and a filtering residual is calculated and reversely fed back to the time-space domain sparse Bayesian optimization algorithm.
  5. 5. The electromagnetic induction-based track traffic vehicle contactless positioning system according to claim 1 is characterized in that a bidirectional fusion and mutual feedback process of two algorithms in a time-space domain calibration feedback module is that an adaptive variable-order Kalman filtering algorithm for multi-source heterogeneous data fusion is based on an optimized space-time data matrix and a noise covariance initial value output by a time-space domain sparse Bayesian optimization algorithm, filtering calibration is completed, filtering residual errors are generated, the filtering residual errors are reversely transmitted to the time-space domain sparse Bayesian optimization algorithm and used for adjusting sparsity constraint coefficients of the time-space domain sparse Bayesian optimization algorithm to optimize space-time data processing precision, and an output result optimized by the time-space domain sparse Bayesian optimization algorithm is input into the adaptive variable-order Kalman filtering algorithm for multi-source heterogeneous data fusion again to iteratively optimize the precision of three-dimensional position coordinates of a calibrated train.
  6. 6. The electromagnetic induction-based contactless positioning system for rail transit vehicles is characterized in that the electromagnetic interference counteracting process of the dynamic magnetic bias compensation module is characterized in that firstly, interference characteristics in a double-path original signal set are extracted through a magnetic interference signal recognition unit, three types of interference including rail steel structure magnetic eddy current interference, peripheral power equipment power frequency interference and magnetic shielding interference of metal parts of the vehicles are accurately recognized, then, magnetic bias signals with the same amplitude, the same frequency and opposite phases as the interference signals are generated through an active magnetic bias generation unit, interference is counteracted through magnetic field superposition, and finally, the amplitude range and the phase reference of double signals are calibrated through a double-signal synchronous compensation unit, so that a calibrated double-signal set is generated.
  7. 7. The electromagnetic induction-based rail transit vehicle contactless positioning system according to claim 1, wherein the distributed magnetostrictive electromagnetic induction array module adopts a 10×8 distributed layout, array units are longitudinally arranged along the running direction of the train and transversely perpendicular to the running direction, each array unit is integrated with a Terfenol-D magnetostrictive alloy sheet, a copper induction coil and a miniature signal acquisition chip, all array units are controlled by a signal synchronous acquisition unit to synchronously acquire two-way original signals, the synchronous error is controlled at microsecond level, and high-frequency impulse noise is primarily shaped and filtered after acquisition.
  8. 8. The electromagnetic induction-based track traffic vehicle contactless positioning system according to claim 1, wherein the track traffic network GIS space-time data acquired by the time-space domain calibration feedback module comprises static space-time data and dynamic space-time data, the static space-time data is track line geographic coordinates, curvature, gradient, stations and turnout positions, the static space-time data is updated once a month, the dynamic space-time data is real-time running speed, acceleration and running direction of a train, the dynamic space-time data is acquired by a vehicle-mounted sensor and uploaded to a trackside signal scheduling center, the track real-time working condition data is acquired by a trackside sensor through 5G wireless transmission, and the track real-time working condition data comprises a track occupation state, a construction section, a magnetic field abnormal section and trackside equipment working state.
  9. 9. The electromagnetic induction-based track traffic vehicle contactless positioning system is characterized in that a weighting fusion process of a vehicle-mounted track side double-end verification module is characterized in that a preset vehicle-mounted end positioning data weight is 0.6, track side real-time position data weight is 0.4, the vehicle-mounted end positioning data are generated by matching three-dimensional position coordinates of a calibrated train with position correction coefficients, track side real-time position data are obtained by acquiring signals by a track side fixed stake unit and then resolving the signals by a track side data processing unit, the track side real-time position data are bound with unique position codes of the track side fixed stake unit, and abnormal values are removed by accuracy verification after the two types of data are weighted and fused to generate a final positioning result.
  10. 10. The contactless positioning system for rail transit vehicles based on electromagnetic induction of claim 9, wherein in the vehicle-mounted trackside double-end verification module, the error feedback value is the fusion deviation of vehicle-mounted end positioning data and trackside real-time position data, and is transmitted to the dynamic magnetic bias compensation module and the double-phase difference solution positioning module through the dedicated feedback bus after being standardized by the error feedback unit, wherein the error feedback value transmitted to the dynamic magnetic bias compensation module is used for adjusting the signal intensity of the active magnetic bias generation unit, and the error feedback value transmitted to the double-phase difference solution positioning module is used for adjusting the space weight coefficient of the array weighted pseudo-range projection algorithm.

Description

Rail transit vehicle contactless positioning system based on electromagnetic induction Technical Field The invention relates to the technical field of rail transit vehicle positioning and electromagnetic induction sensing intersection, in particular to a non-contact positioning system of a rail transit vehicle based on electromagnetic induction. Background The accurate positioning of the rail transit vehicle is a core foundation of train operation scheduling, safety protection and automatic driving, and the contactless positioning gradually replaces the traditional contact positioning scheme because of avoiding mechanical abrasion and adapting to a complex rail environment. The existing non-contact positioning technology is mostly based on single or simple fusion modes such as single electromagnetic induction, GPS/Beidou, inertial navigation and the like, and has a plurality of technical bottlenecks. The traditional electromagnetic induction positioning relies on single induction voltage signal acquisition, is easily influenced by magnetic eddy current of a track steel structure, power frequency interference of peripheral power equipment, magnetic shielding of vehicle metal parts and the like, has poor signal stability, causes larger positioning deviation, adopts conventional Kalman filtering, linear phase difference resolving and other mature algorithms for resolving and calibrating algorithms, is difficult to adapt to scenes of non-uniform distribution of space-time data of track traffic and dynamic change of train speed, and has obvious precision attenuation under shielding working conditions such as high-speed running, tunnel and the like. Meanwhile, the existing scheme generally lacks a full-flow closed-loop calibration mechanism, the solution error of the preamble module cannot be reversely corrected through the subsequent result, accumulated errors are easy to generate, and part of the scheme is additionally provided with a verification link, but only single-end data verification is performed, no vehicle-mounted double-end cooperative verification with the trackside is performed, and local positioning offset errors are difficult to remove. In addition, the perception, resolving and calibration modules of the existing system are mutually independent, the problems of incompatible formats, poor time delay matching and the like exist in data transmission, the overall connection fluency is insufficient, the positioning reliability and the accuracy of a full-speed section and a full working condition cannot be taken into account, and the requirements of rail transit automatic driving and high-density scheduling on sub-meter positioning are difficult to meet. Disclosure of Invention The invention provides a contactless positioning system for rail transit vehicles based on electromagnetic induction, which solves the technical defects of easy interference of positioning precision, weak suitability of working conditions and speeds, lack of an effective closed-loop calibration mechanism, poor module engagement fluency and the like in the conventional contactless positioning technology for rail transit, realizes dynamic closed-loop optimization of train positioning through fusion of architecture, algorithm and sensing modes, and provides stable and reliable positioning support for operation scheduling, safety control and automatic driving of rail transit trains. In order to achieve the above purpose, the invention adopts the following technical scheme: a rail transit vehicle contactless positioning system based on electromagnetic induction, comprising: the distributed magnetostrictive electromagnetic induction array module is used for synchronously collecting magnetostrictive deformation electric signals and electromagnetic induction voltage signals to form a double-path original signal set; The dynamic magnetic bias compensation module is used for receiving the double-path original signal set, identifying and counteracting electromagnetic interference in the double-path original signal set and generating a calibrated double-signal set; The double phase difference resolving and positioning module is internally provided with a complex frequency domain phase unwrapping algorithm and an array weighted pseudo-range projection algorithm, and is used for receiving the calibrated double signal set, and carrying out phase unwrapping and position resolving on the calibrated double signal set through bidirectional fusion and mutual feedback of the two algorithms to generate an initial three-dimensional position coordinate of the train and a double phase difference resolving error value; The time-space domain calibration feedback module is internally provided with a time-space domain sparse Bayesian optimization algorithm and a multi-source heterogeneous data fusion self-adaptive variable order Kalman filtering algorithm, and is used for acquiring the GIS space-time data and the GIS real-time working condition data of the