Search

CN-122015827-A - Multi-source fusion positioning method and device for train, electronic equipment and storage medium

CN122015827ACN 122015827 ACN122015827 ACN 122015827ACN-122015827-A

Abstract

The application discloses a train multisource fusion positioning method, a device, electronic equipment and a storage medium, and relates to the technical field of train multisource fusion positioning. The method comprises the steps of obtaining inertial measurement unit data, wheel speed sensor data, railway operation real-time signal data and preset digital track map data which are collected in a train running process, carrying out discrete division on a current sliding window, constructing state nodes and corresponding state variables, constructing IMU pre-integration factors connected with adjacent state nodes based on the inertial measurement unit data, constructing wheel speed factors of single state nodes based on the wheel speed sensor data, constructing track geometry factors and track occupation factors of the single state nodes based on the inertial measurement unit data, the preset digital track map data and the railway operation real-time signal data, adding each constructed constraint factor into a factor graph, and solving all the state variables to obtain optimal positioning state parameters. The technical scheme improves the positioning precision in the running process of the train.

Inventors

  • DONG SHUNCHANG

Assignees

  • 株洲太昌电子信息技术股份有限公司

Dates

Publication Date
20260512
Application Date
20260210

Claims (10)

  1. 1. The train multisource fusion positioning method is characterized by comprising the following steps of: Acquiring inertial measurement unit data, wheel speed sensor data, railway operation real-time signal data and preset digital track map data which are acquired in a current sliding window in the running process of a train; Dividing discrete time in the current sliding window according to a preset time step, and constructing a state node and a state variable corresponding to the state node for each discrete time, wherein the state variable comprises a train position, a train posture, a train speed and a sensor zero offset; An IMU pre-integral factor connected with adjacent state nodes is built based on the inertial measurement unit data, and a wheel speed factor of a single state node is built based on the wheel speed sensor data; Based on the inertial measurement unit data, preset digital track map data and railway operation real-time signal data, constructing track geometric factors and track occupation factors of single-state nodes; And uniformly adding the IMU pre-integral factor, the wheel speed factor, the track geometry factor and the track occupation factor into the factor graph, and carrying out joint solution on all state variables to obtain optimal positioning state parameters at all moments in a time window, wherein the optimal positioning state parameters comprise position state parameters, attitude state parameters and speed state parameters.
  2. 2. The method of claim 1, wherein constructing the track geometry factor for the single state node based on the inertial measurement unit data, the preset digital track map data, and the railway operation real-time signal data comprises: Determining the train estimated position at each moment in the current time window according to the pre-integration result of the inertial measurement unit data in the current sliding window, and judging the area to which the train belongs according to the train estimated position; selecting a track reference curve corresponding to the region type from the digital track map according to the region type of the region to which the train belongs, and determining a track geometric residual function according to the train estimated position and the track reference curve; And acquiring a measurement noise covariance matrix of the digital orbit map, and determining an orbit geometric factor of a single state node according to the orbit geometric residual function and the measurement noise covariance matrix.
  3. 3. The method according to claim 2, wherein the selecting a track reference curve corresponding to the region type from the digital track map according to the region type of the region to which the train belongs, and determining a track geometry residual function according to the estimated train position and the track reference curve, comprises: If the area of the train is located in the turnout area, acquiring turnout turning-on codes of the turnout area from the railway operation real-time signal data, extracting turnout guide curves of turnout turning-on codes corresponding to turnout turning-on from the digital track map, and calculating the shortest distance from the estimated position of the train to the turnout guide curves as a track geometry residual function; and if the area of the train is located in the non-turnout area, extracting a track center line corresponding to the train estimated position from the digital track map, and calculating the shortest distance from the train estimated position to the track center line as a track geometric residual function.
  4. 4. The method of claim 1, wherein constructing the track occupancy factor for the single state node based on the inertial measurement unit data, the preset digital track map data, and the railway operation real-time signal data comprises: According to the pre-integration result of the inertia measurement unit data in the current sliding window, determining the train estimated position at each moment in the current time window, and acquiring the track occupancy state of the track where the train estimated position is located from the railway operation real-time signal data; When the stock track occupation state is occupied, acquiring the transverse coordinates of the stock track center line of the stock track from the preset digital track map data; determining a transverse constraint residual function according to the transverse coordinates in the train estimated position and the transverse coordinates of the track center line; and acquiring an operation signal noise covariance matrix, and determining the track occupation factor of the single state node according to the transverse constraint residual function and the operation signal noise covariance matrix.
  5. 5. The method according to claim 1, wherein the method further comprises: constructing a discretization error state equation based on the train kinematics model and the sensor error characteristic; Obtaining a zero offset estimated value of a last sliding window sensor, and predicting an error state at the current moment through the discretization error state equation to obtain a prediction error state, wherein the sensor comprises an inertia measurement unit and a wheel speed sensor; Carrying out Kalman updating on the optimal positioning state parameter and the prediction error state to obtain a zero offset estimation correction value of a sensor corresponding to the current moment; correcting the original data acquired by the corresponding sensor according to the zero offset estimation correction value to obtain corrected sensor data, wherein the corrected sensor data is used for a factor graph optimization process in a next sliding window; And carrying out deviation correction on the optimal positioning state parameter according to the zero deviation estimation correction value to obtain a corrected optimal positioning parameter, wherein the corrected optimal positioning parameter is used for train real-time control or navigation.
  6. 6. The method of claim 5, wherein the method further comprises: acquiring environmental data in real time through a vehicle-mounted sensor, and inputting the environmental data into a pre-trained neural network model to obtain a compensation increment predicted value and a corresponding prediction confidence coefficient of an inertial measurement unit; Determining a weighting coefficient of the compensation increment predicted value according to the prediction confidence, and carrying out weighting calculation by combining with a zero offset estimation correction value of an inertia measurement unit to obtain a final compensation increment; And correcting the initial data acquired by the inertial measurement unit and the optimal positioning parameters by using the final compensation increment to obtain the inertial measurement unit data and the optimal positioning parameters after secondary correction.
  7. 7. A multi-source fusion positioning device for a train, comprising: The initial data acquisition module is used for acquiring inertial measurement unit data, wheel speed sensor data, railway operation real-time signal data and preset digital track map data which are acquired in a current sliding window in the running process of the train; The system comprises a state node construction module, a state node detection module and a state node detection module, wherein the state node construction module is used for dividing discrete time in the current sliding window according to a preset time step, and constructing a state node and a state variable corresponding to the state node according to each discrete time, wherein the state variable comprises a train position, a train posture, a train speed and a sensor zero offset; the kinematic constraint factor construction module is used for constructing IMU pre-integral factors connected with adjacent state nodes based on the inertial measurement unit data and constructing wheel speed factors of single state nodes based on the wheel speed sensor data; The track space constraint factor construction module is used for constructing track geometric factors and track occupation factors of single-state nodes based on the inertial measurement unit data, preset digital track map data and railway operation real-time signal data; And the factor graph model solving module is used for uniformly adding the IMU pre-integration factor, the wheel speed factor, the orbit geometric factor and the track occupation factor into the factor graph, and carrying out joint solving on all state variables to obtain optimal positioning state parameters at all moments in a time window, wherein the optimal positioning state parameters comprise position state parameters, attitude state parameters and speed state parameters.
  8. 8. An electronic device, the electronic device comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the train multisource fusion positioning method of any one of claims 1-6.
  9. 9. A computer readable storage medium storing computer instructions for causing a processor to perform the train multisource fusion positioning method according to any one of claims 1-6.
  10. 10. A computer program product, characterized in that the computer program product comprises a computer program which, when executed by a processor, implements a train multisource fusion positioning method according to any one of claims 1-6.

Description

Multi-source fusion positioning method and device for train, electronic equipment and storage medium Technical Field The embodiment of the application relates to the technical field of train positioning, in particular to the technical field of train multi-source fusion positioning, and particularly relates to a method, a device, electronic equipment and a storage medium for train multi-source fusion positioning. Background In a railway locomotive navigation system, the positioning accuracy of a global navigation satellite system is reduced or even completely disabled due to the limited satellite signals such as tunnels, mountain areas, urban canyons, electromagnetic interference and the like, so that the navigation accuracy of train operation is affected. The prior art mainly comprises a loose combination positioning mode of a global navigation satellite system and an inertial measurement unit, wherein the positioning accuracy and reliability are obviously reduced when satellite signals are limited or lose efficacy, a positioning mode of combining a wheel speed sensor and an orbit database is adopted, positioning is mainly carried out by relying on speed integration under the condition of satellite signal failure, errors are easy to accumulate with time to cause the reduction of positioning accuracy, and a point type positioning mode based on a transponder or a beacon is disadvantageous in that a large amount of ground equipment is required to be deployed, the maintenance cost is high, discrete point position information can only be provided, and continuous speed measurement and attitude estimation are difficult to realize. Therefore, how to provide a method for realizing accurate navigation and positioning of a train under a satellite signal limited environment is a technical problem that needs to be solved by those skilled in the art. Disclosure of Invention The application provides a multi-source fusion positioning method, a multi-source fusion positioning device, electronic equipment and a storage medium for a train, so as to improve positioning accuracy in the running process of the train. According to one aspect of the application, a train multisource fusion positioning method is provided, which comprises the following steps: Acquiring inertial measurement unit data, wheel speed sensor data, railway operation real-time signal data and preset digital track map data which are acquired in a current sliding window in the running process of a train; Dividing discrete time in the current sliding window according to a preset time step, and constructing a state node and a state variable corresponding to the state node for each discrete time, wherein the state variable comprises a train position, a train posture, a train speed and a sensor zero offset; An IMU pre-integral factor connected with adjacent state nodes is built based on the inertial measurement unit data, and a wheel speed factor of a single state node is built based on the wheel speed sensor data; Based on the inertial measurement unit data, preset digital track map data and railway operation real-time signal data, constructing track geometric factors and track occupation factors of single-state nodes; And uniformly adding the IMU pre-integral factor, the wheel speed factor, the track geometry factor and the track occupation factor into the factor graph, and carrying out joint solution on all state variables to obtain optimal positioning state parameters at all moments in a time window, wherein the optimal positioning state parameters comprise position state parameters, attitude state parameters and speed state parameters. According to another aspect of the present application, there is provided a train multisource fusion positioning device comprising: The initial data acquisition module is used for acquiring inertial measurement unit data, wheel speed sensor data, railway operation real-time signal data and preset digital track map data which are acquired in a current sliding window in the running process of the train; The system comprises a state node construction module, a state node detection module and a state node detection module, wherein the state node construction module is used for dividing discrete time in the current sliding window according to a preset time step, and constructing a state node and a state variable corresponding to the state node according to each discrete time, wherein the state variable comprises a train position, a train posture, a train speed and a sensor zero offset; the kinematic constraint factor construction module is used for constructing IMU pre-integral factors connected with adjacent state nodes based on the inertial measurement unit data and constructing wheel speed factors of single state nodes based on the wheel speed sensor data; The track space constraint factor construction module is used for constructing track geometric factors and track occupation factors of single-state nodes based on the inertial m