A multi-radar system error correction method based on DBSCAN clustering

Abstract

The application discloses a multi-radar system error correction method based on DBSCAN clustering and relates to the field of radar target tracking. Multi-radar plot fusion can improve target tracking accuracy, but due to the differences between installation and performance parameters of different radars, system errors are inevitably generated, causing the uncertainty of target measurement; meanwhile, under the scenarios of missed detection and false alarm, the system error is difficult to accurately estimate. The method utilizes DBSCAN clustering to classify the offsets between radars (i.e. the vector difference between the extrapolated value of the reference radar measurement and the measurement of the radar to be corrected) within w continuous time stamps, extracts the offset cluster with the highest density and takes the mean value as the real offset, and eliminates the false offsets generated by missed detection and false alarm; meanwhile, filtering the real offset can eliminate random errors and obtain a more accurate system error estimation value.

Description

Technical Field

[0001] This invention relates to a method for correcting errors in multi-radar systems based on DBSCAN clustering in the field of radar technology. Specifically, it is a method for correcting system errors between multiple radar systems based on the DBSCAN (Density-Based Spatial Clustering of Application with Noise) clustering algorithm. Background Technology

[0002] In multi-radar point fusion tracking, it is usually necessary to transform the measurements from each radar into a unified coordinate system before subsequent fusion processing. Due to differences in installation and performance parameters, different radars inevitably introduce systematic errors. These systematic errors cause deviations in the measurements of the same target from different radars within the unified coordinate system, increasing the probability of incorrect point association between multiple radars targeting the same target and affecting target tracking performance. In scenarios without missed detections or false alarms, calculating the average and subtracting the values ​​of measurements from two radars can effectively obtain the systematic error between them. However, in cases with missed detections or false alarms, missed detections prevent the calculation of systematic errors between radars, while false alarms lead to significant deviations in the calculated systematic error. Existing methods struggle to overcome the problem of inaccurate acquisition of systematic errors between multiple radars in scenarios with missed detections or false alarms. Summary of the Invention

[0003] In view of this, the present invention provides a multi-radar system error correction method based on DBSCAN clustering, which can effectively overcome the problem that it is difficult to accurately estimate the error of multi-radar systems under the interference of missed detections and false alarms.

[0004] To meet the above requirements, the technical solution adopted by the present invention is as follows:

[0005] A method for error correction in a multi-radar system based on DBSCAN clustering includes the following steps:

[0006] Step 1: Configure system parameters, including radar ID, radar data transmission cycle, sensor latitude and longitude, measurement error, and detection range;

[0007] Step 2: After the radar transmits the point data, the coordinate system of the radar with the highest data rate is regarded as the reference coordinate system, and the corresponding radar is called the reference radar. Then, based on the latitude, longitude and altitude of the sensors, the measurement values ​​and measurement errors of all radars are converted to Cartesian coordinates, and the measurement errors are compensated to the measurement values ​​to form a measurement dataset in the Cartesian coordinate system, including the measurement of the real target and clutter. The measurement values ​​include radial distance, azimuth angle and elevation angle.

[0008] Step 3, within N consecutive time stamps, the reference radar scans, then carries out association and filtering, if the associated measurement number is greater than the threshold value M, it is considered that the target starting track is successful; after the track starting is successful, the association and filtering operation is continuously carried out until the error remains stable; wherein, M and N are both set values;

[0009] Step 4, a clustering time window w is set, the current time is recorded as t, (t-w+1) ～ The vector difference between the extrapolated value of the non-reference radar and the measurement of the nearest neighbor reference radar within the t time stamp is calculated, and is saved in the offset matrix;

[0010] Step 5, the offset matrix obtained within w time windows is classified based on the DBSCAN clustering method, and each cluster offset class after classification is counted, and the highest data density is extracted as the real offset cluster;

[0011] Step 6, the mean value of all offsets in the real offset cluster at t time is calculated, and filtering processing is carried out, so that the system error between the reference radar and the non-reference radar measurement can be obtained.

[0012] The beneficial effects of the present application are that:

[0013] Due to the differences between installation and performance parameters of different radars, system errors are inevitably generated, causing the uncertainty of target measurement; at the same time, under the scenes of missed detection and false alarm, the system error is difficult to accurately estimate. The present application uses DBSCAN clustering to classify the offsets between radars within w consecutive time stamps, extracts the highest density offset cluster and takes the mean value as the real offset, and eliminates the false offsets generated by missed detection and false alarm; at the same time, filtering the real offset can eliminate random errors and obtain more accurate system error estimation value. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 It is the flow chart of the multi-radar system error correction method of the embodiment of the present application.

[0015] Figure 2 It is the simulation experiment result of the multi-radar (radar 1 and radar 2) system error correction method of the embodiment of the present application.

[0016] Figure 3 It is the simulation experiment result of the multi-radar (radar 1 and radar 3) system error correction method of the embodiment of the present application. DETAILED DESCRIPTION

[0017] The present application will be further described in detail below in combination with the drawings and specific embodiments.

[0018] As Figure 1A multi-radar system error correction method based on DBSCAN clustering, comprising the following steps:

[0019] Step 1, parameter configuration, configure system parameters, including radar ID, radar data sending period, sensor longitude and latitude, measurement error, detection range, etc.

[0020] Step 2, data preprocessing, after the radar sends the track data, the measurement values (radial distance, azimuth angle and pitch angle) of all radars need to be converted into a unified Cartesian coordinate system. Among them, the coordinate system with the highest data rate is usually taken as the reference coordinate system, and the radar is called the reference radar. Then, based on the sensor longitude and latitude, the measurement values and measurement errors of all radars are converted to Cartesian coordinates, and the measurement errors are compensated to the measurement values to form a measurement data set (including real target measurement and clutter) in the Cartesian coordinate system.

[0021] Step 3, track initiation and maintenance, use the measurement of the reference radar to initiate a multi-target track, and use the M / N logic method to initiate and terminate the track. The process is as follows: within N consecutive time stamps of the reference radar scanning, if the number of associated measurements is greater than the threshold value M, it is considered that the target track initiation is successful; after the track initiation is successful, continue to perform association and filtering operations until the error remains stable; wherein M and N are both set values.

[0022] Step 4, multi-window offset calculation, set the clustering time window w, the current time is t, calculate the vector difference between the extrapolated value of the non-reference radar and its nearest neighbor reference radar measurement within (t-w+1) ～ t time stamp, and save it to the offset matrix.

[0023] Step 5, DBSCAN clustering, based on the DBSCAN clustering method, the offset matrix obtained within w time windows is automatically classified, and the classified offset classes are counted, and the highest data density is extracted as the real offset cluster.

[0024] Step 6, system error calculation, calculate the mean value of all offsets in the real offset cluster at time t, and perform first-order extended Kalman filtering on the average offset, which can obtain the system error between the reference radar and the non-reference radar measurement.

[0025] Method simulation experiment verification:

[0026] Scenario simulation: 3 targets, 3 radars, the scanning period of radar 1 is 1.0s, the latitude and longitude height is (28.4, 119.2, 100), the radar installation angle error is (0.1°, 0.2°, -0.1°), and the distance error is (10m, 5m, 5m). The scanning period of radar 2 is 1.2s, the latitude and longitude height is (28.4, 120.1, 706.6), the radar installation angle error is (-0.1°, 0.2°, -0.1°), and the distance error is (-10m, -5m, 5m). The scanning period of radar 3 is 1.3s, the latitude and longitude height is (28.8, 119.7, 404.1), the radar installation angle error is (-0.1°, -0.2°, -0.1°), and the distance error is (-10m, 5m, -5m). Radar 1 is the reference radar, and radars 2 and 3 are non-reference radars. In addition, a clutter is randomly generated around the true measurement of each target in each timestamp, and the radar measurement miss detection rate is set to 0.1.

[0027] The system error calculation results between the reference radar and the non-reference radar are shown in Figure 2 and Figure 3 From the figure, it can be seen that due to the installation error, there is a difference between the system error results of different radars. However, the specific value of the system error in each timestamp can be quantitatively calculated by using the method.

[0028] In order to evaluate the accuracy of the system error calculation, the measurement values processed by the method of the application and the measurement values not processed by the method of the application are subjected to track fusion (i.e. association and filtering), and the filtering results are compared. The specific filtering errors are shown in the following table:

[0029] Constant velocity straight line motion Target 1 filtered RMSE Target 2 filtered RMSE Target 3 filtered RMSE Corrected system error 23.18 21.62 23.10 Uncorrected system error 229.86 222.13 221.53

[0030] As can be seen from the above table, the target tracking effect is significantly improved after the measurement values processed by the method of the application are subjected to track fusion, and it can be directly seen that the filtering error is much smaller than the filtering result of the uncorrected system error, which shows that the system error value between radars can be calculated more accurately by the application.

[0031] In summary, the system error offset in multiple timestamps can be clustered by DBSCAN, the problem that the system error is difficult to estimate due to miss detection and false alarm can be better solved, the system error can be corrected correctly, the real-time tracking accuracy of the target is significantly improved, and the application prospect is good.

Claims

1. A multi-radar system error correction method based on DBSCAN clustering, characterized in that, Comprising the following steps: Step 1, configure system parameters, including radar ID, radar data sending period, sensor longitude and latitude, measurement error and detection range; Step 2, after the radar sends the track data, the coordinate system with the highest data rate is regarded as the reference coordinate system, and the corresponding radar is called the reference radar; Then, based on the sensor longitude and latitude, the measurement values and measurement errors of all radars are converted to Cartesian coordinates, and the measurement errors are compensated to the measurement values to form a measurement data set in the Cartesian coordinate system, including real target measurements and clutter; wherein the measurement values include radial distance, azimuth angle and elevation angle; Step 3, within N consecutive time stamps, the reference radar is scanned, then associated and filtered, and if the number of associated measurements is greater than the threshold value M, it is considered as a successful target starting track; after the track starting is successful, the association and filtering operation is continued until the error remains stable; wherein M and N are both set values; Step 4, set the clustering time window w, the current time is denoted as t, calculate (t-w+1) ～ The vector difference between the non-reference radar measurement in the t timestamp and the extrapolated value of the measurement of the nearest neighbor reference radar, and save it to the offset matrix. Step 5, based on the DBSCAN clustering method, the offset matrix obtained within the w time window is classified, and the offset class of each cluster after classification is counted, and the data density highest is extracted as the real offset cluster; Step 6, calculate the mean value of all offsets in the real offset cluster at time t, and filter the average offset, that is, the system error between the reference radar and the non-reference radar measurement can be obtained.

Images