Metro train speed measurement method based on millimeter wave radar

Abstract

The invention discloses a metro train speed measurement method based on a millimeter-wave radar, and the method comprises the steps: S1, millimeter-wave radar equipment installation and calibration: enabling the millimeter-wave radar to be fixedly installed at the outer side of a train cab through a tool clamp, testing and measuring the installation posture of the millimeter-wave radar, and employing a gyro machine as the basis of dynamic calibration; step S2, radar signal acquisition and processing: millimeter wave radar message data are analyzed, interference of a subway line environment is reduced from the aspects of radar working parameter tuning and beam radiation optimization, and conditions are set to trigger high-frequency data acquisition; s3, data preprocessing and data analysis: preprocessing the collected data from the aspects of noise values, outliers and data reduction, identifying a detection target by adopting a density-based spatial clustering algorithm, and outputting target point cloud data; and S4, calculating the train speed and outputting the train speed in real time. According to the invention, the millimeter-wave radar is installed and the detected millimeter-wave radar point cloud data is analyzed in real time, so that accurate measurement of the train running speed is realized.

Description

Technical Field

[0001] The invention relates to urban rail transit train speed measurement and the technical field of millimeter wave radar, and in particular to a subway train speed measurement method based on millimeter wave radar. Background Art

[0002] Train speed measurement is the core function of the rail transit signal system and an important basis for ensuring the safe and efficient operation of trains. The current CBTC system mainly uses axle speed sensors to determine the speed of the train. It is based on the principle of pulse speed sensors and measures the rotation speed of the train wheels to infer the running speed of the train. This method has the advantages of simple structure, high measurement accuracy, and relatively low cost. However, in rainy and snowy weather or uphill conditions, the train wheels are prone to slip or idle. The train speed and distance data values ​​calculated by the axle sensors deviate greatly from the actual working conditions, which may even affect driving safety in severe cases. With the development of unmanned driving technology for subway trains, rail transit signal systems have put forward higher requirements for train speed measurement accuracy and anti-interference capabilities.

[0003] Millimeter waves are electromagnetic waves between microwaves and light waves, with wavelengths ranging from 1 to 10 mm and frequencies ranging from 30 to 300 GHz. They have the advantages of long detection distance, strong penetration and anti-interference capabilities, and have been widely used in the field of industrial manufacturing. For example, in the field of autonomous driving, millimeter wave radars can provide all-weather and all-day environmental perception data, support the decision-making and planning of autonomous driving vehicles, and provide a basis for vehicle obstacle detection and avoidance, path planning and decision-making, blind spot monitoring and other functions.

[0004] In the field of rail transit, patent CN218974594U proposes a radar speed measurement system for rail transit. This solution uses a vehicle-mounted millimeter-wave radar component to send and receive digital signals, and is connected to the vehicle-mounted control system through a signal processing, control and transmission module for transmitting and receiving electromagnetic waves. This solution focuses on millimeter-wave radar hardware components and power conversion, but does not provide a method for processing the collected radar signal data. Patent CN110058239B provides a vehicle-mounted millimeter-wave radar device and a target detection method. Through a device composed of a radio frequency front end, a power supply, and a signal processing module, it transmits and receives target echo signals, pre-processes the received data, and filters out the fences on both sides of the track. It provides a method for detecting targets, but there is no method for determining the speed of a dynamically moving train.

[0005] In the field of vehicle-mounted speed measurement applications, patent CN109375202B proposes a method for measuring the speed of external vehicles based on vehicle-mounted millimeter-wave radar. The beat signals of the triangular band and the sawtooth band in the mixed wave are Fourier transformed respectively, and the distance information calculated by the two bands is compared to obtain the distance and speed information of the external vehicle target. This method has problems with clutter filtering and false target interference. Patent CN112946623B installs a 77GHz millimeter-wave radar on the vehicle, calculates the distance, angle, speed, signal strength and other information of the target according to the echo signal, constructs a point cloud data set and screens out valid point clouds to determine the speed of the first vehicle behind itself. This solution is used to determine the speed of the rear vehicle, and lacks a method for measuring the speed of the vehicle itself. Patent CN111308458B proposes a vehicle speed estimation method based on vehicle-mounted millimeter-wave radar. The radar is installed in the horizontal position forward of the vehicle to be tested. By receiving the forward target reflected wave signal, the relative speed of the target in the field of view is analyzed to evaluate the vehicle's own speed and perform correlation filtering to obtain the vehicle's own speed estimation result with confidence. This method provides real-time and accurate estimation of the vehicle's own speed, but the error threshold, filtering parameters, etc. rely on prior knowledge or are based on a large number of experiments.

[0006] In view of the defects and shortcomings in the prior art, a method for measuring the speed of a subway train based on millimeter-wave radar is provided. Summary of the invention

[0007] The purpose of the present invention is to overcome the existing defects and provide a subway train speed measurement method based on millimeter wave radar, which realizes accurate measurement of the train running speed by installing a vehicle-mounted millimeter wave radar and analyzing the detected millimeter wave radar point cloud data in real time.

[0008] The technical solution to achieve the above purpose is:

[0009] A method for measuring subway train speed based on millimeter wave radar, comprising:

[0010] Step S1, installation and calibration of millimeter wave radar equipment: the millimeter wave radar is fixedly installed on the outside of the train cab by a fixture, and its installation posture is experimentally measured, and a gyroscope is used as a basis for dynamic calibration;

[0011] Step S2, radar signal acquisition and processing: parsing millimeter wave radar message data, reducing interference from the subway line environment by optimizing radar operating parameters and beam radiation, and setting conditions to trigger high-frequency data acquisition;

[0012] Step S3, data preprocessing and data analysis: preprocess the collected data from the aspects of noise value, outlier value and data reduction, and use density-based spatial clustering algorithm to identify the detection target and output the target point cloud data;

[0013] Step S4, calculate the train speed and output it in real time: convert the point cloud data detected by the millimeter-wave radar into target data in the train coordinate system, determine the validity of multiple frames of continuous millimeter-wave radar point cloud data, and output the effective train speed.

[0014] Preferably, in step S1, the millimeter wave radar device uses a 77 GHz millimeter wave radar, which is a 4-transmit 12-receive flat-panel array antenna;

[0015] During the static calibration process, the radar corner reflector tool is set up at different positions on the line. By analyzing the detection data output by the radar, the initial azimuth, pitch angle and roll angle of the radar relative to the train installation are obtained;

[0016] The lateral and longitudinal offset distances between the center of the radar array and the center of the train section are measured and used as the initial installation calibration value of the radar and as the basis for the next step of dynamic calibration calculation; the result of the static calibration is expressed as follows:

[0017]

[0018] In the formula, is the initial azimuth angle of the center of the millimeter-wave radar array relative to the center of the train section, is the initial pitch angle of the center of the millimeter-wave radar array relative to the center of the train section, is the initial roll angle of the center of the millimeter-wave radar array relative to the center of the train section, is the horizontal offset of the center of the millimeter-wave radar array relative to the center of the train section, is the longitudinal axis offset of the center of the millimeter wave radar array relative to the center of the train section, is the vertical offset of the center of the millimeter-wave radar array relative to the center of the train section;

[0019] Since the car body and the axle bogie are non-rigidly connected, the attitude of the radar relative to the train is also changing during the dynamic driving of the train, that is, the radar attitude needs to be dynamically calibrated;

[0020] The static attitude calibration value is corrected by measuring the attitude angle of the gyroscope rigidly mounted on the train in real time. The train position during static calibration is taken as the gyroscope attitude zero position. The vehicle attitude angle measured by the gyroscope at time t is:

[0021]

[0022] Using this value to correct the static calibration attitude data, the attitude calibration value of the train at time t can be obtained as follows:

[0023]

[0024] in,

[0025]

[0026] In the formula, is the azimuth angle of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the pitch angle of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the rolling angle of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the horizontal displacement of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the longitudinal axis offset of the center of the millimeter wave radar array relative to the center of the train section at time t, It is the vertical offset of the center of the millimeter-wave radar array relative to the center of the train section at time t.

[0027] Preferably, in step S2, the data output by the millimeter wave radar CAN interface is parsed to obtain the lateral X distance, longitudinal Y distance, point velocity V, azimuth A and RCS value of the radar detection target point according to the defined message format;

[0028] The Doppler frequency shift effect generated by high-speed subway trains is used to optimize the radar's transmission frequency and receiving bandwidth, and dynamically adjust the millimeter-wave radar's operating mode, transmission power, output target range, and target RCS value interval parameters;

[0029] Set enhanced acquisition conditions. When the train speed is judged to be lower than the threshold, it will automatically switch to high transmission power and high receiving sensitivity mode.

[0030] For specific scenarios of slipping, sudden acceleration or deceleration, the real-time acceleration and slip status of the train are acquired and monitored from the signal system, and trigger conditions are set. If the train acceleration threshold or wheel-rail friction exceeds the threshold, a higher frequency data acquisition mode is started to capture more detailed information.

[0031] Preferably, in step S2, if the train acceleration threshold or the wheel-rail friction exceeds the threshold, the millimeter wave radar data acquisition frequency is adjusted; if it does not exceed the threshold, the data is collected at the default operating frequency.

[0032] Preferably, in step S3, the preprocessing includes:

[0033] In terms of noise value processing:

[0034] Filter out point data that obviously deviates from the orbital environment through target RCS value, target distance, and target Doppler velocity value conditions;

[0035] In terms of outlier processing, the method based on distance nearest neighbor is adopted:

[0036] For each data point in the data within the sampling time segment, calculate the distance between it and all other data points, and find its k nearest neighbors for each data point;

[0037] Calculate the average distance of each data point to its k nearest neighbors, and compare the average distance with the set threshold. If it is greater than the set threshold, the data is considered an outlier. The basic formula for calculating the distance between millimeter-wave radar data point a (x1, y1) and point b (x2, y2) is:

[0038]

[0039] In terms of data specification:

[0040] Taking 100ms as the time window, the data with similar attributes within the time window are reconstructed to generate a smaller data set.

[0041] Preferably, in step S3, a density-based spatial clustering algorithm is used to identify the detection target, and outputting the target point cloud data includes:

[0042] The pre-processed millimeter-wave radar detection target point data of each frame is constructed into a sample set, and the clustering parameters are adjusted according to the track line characteristics: clustering radius eps (ε) and the minimum number of cluster points minPts;

[0043] The originally chaotic and discrete millimeter-wave radar point data is converted into target point cloud data. The position of the point cloud cluster center is regarded as the target point position of the frame data, and the average value of the eigenvalue of the point cloud is regarded as the target point eigenvalue of the frame data.

[0044] Preferably, in step S4, the target coordinate system conversion includes:

[0045] The installation posture of the radar is determined through the step S1 process:

[0046]

[0047] The coordinates of the cluster point cloud i output by the millimeter-wave radar detection in the radar coordinate system at time frame t are:

[0048]

[0049] The cluster point cloud i can be converted into a value in the train coordinate system by the following conversion relationship:

[0050]

[0051] Similarly, the velocity value of the clustered point cloud i at time t in the millimeter-wave radar coordinate system is converted to the velocity value in the geographic coordinate system according to the following formula:

[0052]

[0053] Since the subway train runs on the track, the longitudinal axis of the train is always facing the direction of the track. Therefore, in the above formula, To represent the velocity component in the direction of the train’s forward motion, It is the vertical velocity component that characterizes the train on the slope.

[0054] Preferably, in step S4, continuous multi-frame point cloud data is used to estimate the train speed, that is:

[0055] The train speed calculation method at time frame t is:

[0056]

[0057] Continuously store the clustered point cloud data of the time frame t-2 and the time frame t-1, and calculate the train speed of the above time frames;

[0058] If the difference Δ2-Δ1 between the train speed change Δ1 between the frame at time t and the frame at time t-1 and the train speed change Δ2 between the frame at time t-2 and the frame at time t-1 is within the set threshold range;

[0059] Then the train speed at time frame t is valid, and the speed value is output;

[0060] Otherwise, let the train speed at time t-1 plus the speed difference Δ2-Δ1 of the previous two frames be the train speed at time t.

[0061] The beneficial effects of the present invention are as follows: the present invention adopts millimeter-wave radar as a sensor for measuring the speed of subway trains. The equipment is simple to install, free from the installation constraints of bogies and axles, and easy to maintain and update; the principle of millimeter-wave radar is the Doppler effect, and the speed value can be sensitively output when there is relative motion between the train and the track, avoiding the influence of idling and slipping of the train wheel hub; the scheme further reduces the interference of the rail transit line environment on millimeter-wave radar detection from the aspects of hardware parameter tuning and data processing, and can improve the accuracy of train speed measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a flow chart of a method for measuring subway train speed based on millimeter wave radar according to the present invention;

[0063] Figure 2 It is a schematic diagram of the installation of the millimeter wave radar in the present invention. DETAILED DESCRIPTION

[0064] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0065] The present invention will be further described below in conjunction with the accompanying drawings.

[0066] The existing CBTC signal system uses wireless communication and high-precision positioning technology to achieve safe communication and automatic control between trains. As an important part of the CBTC system, train speed measurement provides the system with real-time and accurate train running speed.

[0067] At present, the train speed measurement is mainly realized by axle speed sensors, which are usually installed on an axle of the train bogie. As the train wheels rotate, a certain number of pulse square waves are output. Combined with data such as wheel circumference, the speed of the train can be determined. This method is the basic means of train speed measurement and is simple and reliable. However, it is also easily affected by factors such as wheel idling and slipping, which affects the accuracy of speed measurement.

[0068] like Figure 1 As shown, a method for measuring subway train speed based on millimeter wave radar includes:

[0069] Step S1, installation and calibration of millimeter-wave radar equipment: The millimeter-wave radar is fixedly installed on the outside of the train cab by means of a fixture, and its installation posture is experimentally measured, using a gyroscope as the basis for dynamic calibration.

[0070] In the embodiment, the line characteristics of urban rail transit, such as elevated sections, tunnel sections, and ground sections, as well as rail environmental factors such as isolation fences, utility poles, and distribution boxes are taken into consideration. The radar beam range in the horizontal and elevation directions in this scenario is analyzed, and the millimeter wave radar equipment uses a 77GHz millimeter wave radar, which is a 4-transmit 12-receive flat-panel array antenna; the horizontal and vertical widths of the beam have different performances at close range and long range. In order to make full use of the radar signal characteristics, the radar is fixedly installed on the outside of the train driver's room through a customized fixture, such as Figure 2 As shown, the array plane maintains a specified angle with the train section to achieve the best detection field of view. Static calibration and dynamic calibration are combined to analyze the millimeter wave attitude information in real time during the train's travel.

[0071] During the static calibration process, the radar corner reflector tool is set up at different positions on the line. By analyzing the detection data output by the radar, the initial azimuth, pitch angle and roll angle of the radar relative to the train installation are obtained;

[0072] The lateral and longitudinal offset distances between the center of the radar array and the center of the train section are measured and used as the initial installation calibration value of the radar and as the basis for the next step of dynamic calibration calculation; the result of the static calibration is expressed as follows:

[0073]

[0074] In the formula, is the initial azimuth angle of the center of the millimeter-wave radar array relative to the center of the train section, is the initial pitch angle of the center of the millimeter-wave radar array relative to the center of the train section, is the initial roll angle of the center of the millimeter-wave radar array relative to the center of the train section, is the horizontal offset of the center of the millimeter-wave radar array relative to the center of the train section, is the longitudinal axis offset of the center of the millimeter wave radar array relative to the center of the train section, is the vertical offset of the center of the millimeter-wave radar array relative to the center of the train section;

[0075] Since the car body and the axle bogie are non-rigidly connected, the attitude of the radar relative to the train is also changing during the dynamic driving of the train, that is, the radar attitude needs to be dynamically calibrated;

[0076] The static attitude calibration value is corrected by measuring the attitude angle of the gyroscope rigidly mounted on the train in real time. The train position during static calibration is taken as the gyroscope attitude zero position. The vehicle attitude angle measured by the gyroscope at time t is:

[0077]

[0078] Using this value to correct the static calibration attitude data, the attitude calibration value of the train at time t can be obtained as follows:

[0079]

[0080] in,

[0081]

[0082] In the formula, is the azimuth angle of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the pitch angle of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the rolling angle of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the horizontal displacement of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the longitudinal axis offset of the center of the millimeter wave radar array relative to the center of the train section at time t, It is the vertical offset of the center of the millimeter-wave radar array relative to the center of the train section at time t.

[0083] Step S2, radar signal acquisition and processing: parsing millimeter wave radar message data, reducing interference from the subway line environment by tuning radar operating parameters and optimizing beam radiation, and setting conditions to trigger high-frequency data acquisition.

[0084] In the embodiment, the data output by the millimeter wave radar CAN interface is parsed, and the lateral X distance, longitudinal Y distance, point speed V, azimuth A and RCS value of the radar detection target point are parsed according to the defined message format;

[0085] For rail transit application scenarios, due to the narrow and long subway lines, the ground section is affected by trackside facilities, isolation fences, and electromagnetic interference from the open environment; the underground section is closed, and the tunnel wall will produce a multipath effect on the radar signal, thereby forming an interference target output. In order to reduce the impact of interference targets on the speed measurement value, and considering the unique impact of the train's own movement on the radar signal data characteristics, the Doppler frequency shift effect generated by the high-speed operation of the subway train is used to optimize the radar's transmission frequency and receiving bandwidth, and the millimeter-wave radar working mode, transmission power, output target range, and target RCS value interval parameters are dynamically adjusted;

[0086] At the same time, in view of the limitation of millimeter-wave radar's insufficient resolution when the train is running at low speed, enhanced acquisition conditions are set. When the train speed is judged to be lower than the threshold, it automatically switches to high transmission power and high receiving sensitivity mode, so that the train can send a signal of sufficient strength even when running at low speed;

[0087] For specific scenarios of slipping, sudden acceleration or deceleration, the real-time acceleration and slip status of the train are obtained and monitored from the signal system, and the trigger conditions are set. If the train acceleration threshold or wheel-rail friction exceeds the threshold, a higher frequency data acquisition mode is started to capture more detailed information; if the train acceleration threshold or wheel-rail friction exceeds the threshold, the millimeter wave radar data acquisition frequency is adjusted. If it does not exceed the threshold, the data is collected at the default operating frequency.

[0088] Step S3, data preprocessing and data analysis: preprocess the collected data from the aspects of noise value, outlier value and data reduction, and use density-based spatial clustering algorithm to identify the detection target and output the target point cloud data.

[0089] In an embodiment, the pretreatment comprises:

[0090] In terms of noise value processing:

[0091] Filter out point data that obviously deviates from the orbital environment through target RCS value, target distance, and target Doppler velocity value conditions;

[0092] In terms of outlier processing, the method based on distance nearest neighbor is adopted:

[0093] For each data point in the data within the sampling time segment, calculate the distance between it and all other data points, and find its k nearest neighbors for each data point;

[0094] Calculate the average distance of each data point to its k nearest neighbors, and compare the average distance with the set threshold. If it is greater than the set threshold, the data is considered an outlier. The basic formula for calculating the distance between millimeter-wave radar data point a (x1, y1) and point b (x2, y2) is:

[0095]

[0096] In terms of data specification:

[0097] Taking 100ms as the time window, the data with similar attributes within the time window is reconstructed to generate a smaller data set, reducing the computational complexity of subsequent online data analysis.

[0098] In the embodiment, a density-based spatial clustering algorithm is used to identify the detection target, and the output target point cloud data includes:

[0099] The pre-processed millimeter-wave radar detection target point data of each frame is constructed into a sample set, and the clustering parameters are adjusted according to the track line characteristics: clustering radius eps (ε) and the minimum number of cluster points minPts;

[0100] The originally chaotic and discrete millimeter-wave radar point data is converted into target point cloud data. The position of the point cloud cluster center is regarded as the target point position of the frame data, and the average value of the eigenvalue of the point cloud is regarded as the target point eigenvalue of the frame data.

[0101] Step S4, calculate the train speed and output it in real time: convert the point cloud data detected by the millimeter-wave radar into target data in the train coordinate system, determine the validity of multiple frames of continuous millimeter-wave radar point cloud data, and output the effective train speed.

[0102] In the embodiment, the target coordinate system conversion includes:

[0103] The installation posture of the radar is determined through step S1:

[0104]

[0105] The coordinates of the cluster point cloud i output by the millimeter-wave radar detection in the radar coordinate system at time frame t are:

[0106]

[0107] The cluster point cloud i can be converted into a value in the train coordinate system by the following conversion relationship:

[0108]

[0109] Similarly, the velocity value of the clustered point cloud i at time t in the millimeter-wave radar coordinate system is converted to the velocity value in the geographic coordinate system according to the following formula:

[0110]

[0111]

[0112] Since the subway train runs on the track, the longitudinal axis of the train is always facing the direction of the track. Therefore, in the above formula, To represent the velocity component in the direction of the train’s forward motion, It is the vertical velocity component that characterizes the train on the slope.

[0113] In the embodiment, continuous multi-frame point cloud data is used to estimate the train speed, that is:

[0114] The train speed calculation method at time frame t is:

[0115]

[0116] Continuously store the clustered point cloud data of the time frame t-2 and the time frame t-1, and calculate the train speed of the above time frames;

[0117] If the difference Δ2-Δ1 between the train speed change Δ1 between the frame at time t and the frame at time t-1 and the train speed change Δ2 between the frame at time t-2 and the frame at time t-1 is within the set threshold range;

[0118] Then the train speed at time frame t is valid, and the speed value is output;

[0119] Otherwise, let the train speed at time t-1 plus the speed difference Δ2-Δ1 of the previous two frames be the train speed at time t.

[0120] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some or all of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for measuring subway train speed based on millimeter wave radar, characterized in that: include: Step S1, installation and calibration of millimeter wave radar equipment: the millimeter wave radar is fixedly installed on the outside of the train cab by a fixture, and its installation posture is experimentally measured, and a gyroscope is used as a basis for dynamic calibration; Step S2, radar signal acquisition and processing: parsing millimeter wave radar message data, reducing interference from the subway line environment by optimizing radar operating parameters and beam radiation, and setting conditions to trigger high-frequency data acquisition; Step S3, data preprocessing and data analysis: preprocess the collected data from the aspects of noise value, outlier value and data reduction, and use density-based spatial clustering algorithm to identify the detection target and output the target point cloud data; Step S4, calculate the train speed and output it in real time: convert the point cloud data detected by the millimeter-wave radar into target data in the train coordinate system, determine the validity of multiple frames of continuous millimeter-wave radar point cloud data, and output the effective train speed.

2. The method for measuring subway train speed based on millimeter wave radar according to claim 1, characterized in that: In step S1, the millimeter wave radar device uses a 77 GHz millimeter wave radar, which is a 4-transmit 12-receive flat-panel array antenna; During the static calibration process, the radar corner reflector tool is set up at different positions on the line. By analyzing the detection data output by the radar, the initial azimuth, pitch angle and roll angle of the radar relative to the train installation are obtained; The lateral and longitudinal offset distances between the center of the radar array and the center of the train section are measured and used as the initial installation calibration value of the radar and as the basis for the next step of dynamic calibration calculation; the result of the static calibration is expressed as follows: In the formula, is the initial azimuth angle of the center of the millimeter-wave radar array relative to the center of the train section, is the initial pitch angle of the center of the millimeter-wave radar array relative to the center of the train section, is the initial roll angle of the center of the millimeter-wave radar array relative to the center of the train section, is the horizontal offset of the center of the millimeter-wave radar array relative to the center of the train section, is the longitudinal axis offset of the center of the millimeter wave radar array relative to the center of the train section, is the vertical offset of the center of the millimeter-wave radar array relative to the center of the train section; Since the car body and the axle bogie are non-rigidly connected, the attitude of the radar relative to the train is also changing during the dynamic driving of the train, that is, the radar attitude needs to be dynamically calibrated; The static attitude calibration value is corrected by measuring the attitude angle of the gyroscope rigidly mounted on the train in real time. The train position during static calibration is taken as the gyroscope attitude zero position. The vehicle attitude angle measured by the gyroscope at time t is: Using this value to correct the static calibration attitude data, the attitude calibration value of the train at time t can be obtained as follows: in, In the formula, is the azimuth angle of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the pitch angle of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the rolling angle of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the horizontal displacement of the center of the millimeter-wave radar array relative to the center of the train section at time t, is the longitudinal axis offset of the center of the millimeter wave radar array relative to the center of the train section at time t, It is the vertical offset of the center of the millimeter-wave radar array relative to the center of the train section at time t.

3. The method for measuring subway train speed based on millimeter wave radar according to claim 1, characterized in that: In the step S2, the data output by the millimeter wave radar CAN interface is parsed to obtain the lateral X distance, longitudinal Y distance, point speed V, azimuth A and RCS value of the radar detection target point according to the defined message format; The Doppler frequency shift effect generated by high-speed subway trains is used to optimize the radar's transmission frequency and receiving bandwidth, and dynamically adjust the millimeter-wave radar's operating mode, transmission power, output target range, and target RCS value interval parameters; Set enhanced acquisition conditions. When the train speed is judged to be lower than the threshold, it will automatically switch to high transmission power and high receiving sensitivity mode. For specific scenarios of slipping, sudden acceleration or deceleration, the real-time acceleration and slip status of the train are acquired and monitored from the signal system, and trigger conditions are set. If the train acceleration threshold or wheel-rail friction exceeds the threshold, a higher frequency data acquisition mode is started to capture more detailed information.

4. The method for measuring subway train speed based on millimeter wave radar according to claim 3, characterized in that: In step S2, if the train acceleration threshold or the wheel-rail friction exceeds the threshold, the millimeter wave radar data acquisition frequency is adjusted; if it does not exceed the threshold, the data is collected at the default operating frequency.

5. The method for measuring subway train speed based on millimeter wave radar according to claim 1, characterized in that: In step S3, the preprocessing includes: In terms of noise value processing: Filter out point data that obviously deviates from the orbital environment through target RCS value, target distance, and target Doppler velocity value conditions; In terms of outlier processing, the method based on distance nearest neighbor is adopted: For each data point in the data within the sampling time segment, calculate the distance between it and all other data points, and find its k nearest neighbors for each data point; Calculate the average distance of each data point to its k nearest neighbors, and compare the average distance with the set threshold. If it is greater than the set threshold, the data is considered an outlier. The basic formula for calculating the distance between millimeter-wave radar data point a (x1, y1) and point b (x2, y2) is: In terms of data specification: Taking 100ms as the time window, the data with similar attributes within the time window are reconstructed to generate a smaller data set.

6. The method for measuring subway train speed based on millimeter wave radar according to claim 1, characterized in that: In step S3, a density-based spatial clustering algorithm is used to identify the detection target, and outputting the target point cloud data includes: The pre-processed millimeter-wave radar detection target point data of each frame is constructed into a sample set, and the clustering parameters are adjusted according to the track line characteristics: clustering radius eps (ε) and the minimum number of cluster points minPts; The originally chaotic and discrete millimeter-wave radar point data is converted into target point cloud data. The position of the point cloud cluster center is regarded as the target point position of the frame data, and the average value of the eigenvalue of the point cloud is regarded as the target point eigenvalue of the frame data.

7. The method for measuring subway train speed based on millimeter wave radar according to claim 2, characterized in that: In step S4, the target coordinate system conversion includes: The installation posture of the radar is determined through the step S1 process: The coordinates of the cluster point cloud i output by the millimeter-wave radar detection in the radar coordinate system at time frame t are: The cluster point cloud i can be converted into a value in the train coordinate system by the following conversion relationship: Similarly, the velocity value of the clustered point cloud i at time t in the millimeter-wave radar coordinate system is converted to the velocity value in the geographic coordinate system according to the following formula: Since the subway train runs on the track, the longitudinal axis of the train is always facing the direction of the track. Therefore, in the above formula, To represent the velocity component in the direction of the train’s forward motion, It is used to characterize the vertical velocity component of the train on a slope.

8. The method for measuring subway train speed based on millimeter wave radar according to claim 7, characterized in that: In step S4, continuous multi-frame point cloud data is used to estimate the train speed, that is: The train speed calculation method at time frame t is: Continuously store the clustered point cloud data of the time frame t-2 and the time frame t-1, and calculate the train speed of the above time frames; If the difference Δ2-Δ1 between the train speed change Δ1 between the frame at time t and the frame at time t-1 and the train speed change Δ2 between the frame at time t-2 and the frame at time t-1 is within the set threshold range; Then the train speed at time frame t is valid, and the speed value is output; Otherwise, let the train speed at time t-1 frame plus the speed difference Δ2-Δ1 of the previous two frames be the train speed at time t.

Images