System error correction method and system based on ECEF and Q-type clustering analysis

Abstract

The invention discloses a three-coordinate system error correction method based on ECEF and Q-type clustering analysis. Aiming at the problem that multiple nodes cannot cooperatively track due to system errors of sensors in a multi-sensor target tracking system, the method comprises the following steps: firstly, collecting multi-source data, preprocessing the data, then fusing the multi-source data, transferring system error information obtained after fusion to an ECEF coordinate system, and uniformly expressing and processing the system error information; and performing Q-type clustering analysis on the converted system error data, and correcting the system error according to a Q-type clustering result, thereby improving the accuracy and consistency of the data.

Description

Technical Field

[0001] The invention belongs to the field of system error correction in multi-source data fusion, and specifically relates to a system error correction method and system based on ECEF and Q-type cluster analysis. Background Art

[0002] Radar multi-source data fusion is a key information processing technology that aims to improve the performance and accuracy of radar systems by integrating data from different radar systems. However, in practical applications, there are some systematic errors in radar systems, which will affect the accuracy and fusion effect of data. Therefore, correcting the systematic errors in radar multi-source data fusion has become an important research direction.

[0003] Radar system errors are caused by many factors, including hardware and environment. For example, the non-ideal characteristics of the radar antenna, the noise of the receiver, the attenuation of signal transmission, and the influence of atmospheric factors can all lead to measurement errors in the radar system. These errors will have a negative impact on the accuracy and reliability of radar data, thereby affecting the results of multi-source radar data fusion.

[0004] Therefore, further research and exploration of the systematic error correction methods and techniques in radar multi-source data fusion, as well as their effects and limitations in practical applications, have important theoretical and practical significance. By correcting the systematic error, the accuracy and consistency of radar data can be improved, further promoting the development and application of radar multi-source data fusion technology. However, traditional systematic error correction methods may have problems such as positioning errors and data inconsistency, which affect the accuracy and reliability of the fusion results. Summary of the invention

[0005] In order to solve the above technical problems and improve the accuracy of system error estimation in multi-source data fusion, the present invention proposes a system error correction method based on ECEF and Q-type cluster analysis.

[0006] The technical solution for achieving the purpose of the present invention is: a system error correction method based on ECEF and Q-type cluster analysis, comprising:

[0007] Step 1: Collect measurement data containing systematic errors from different data sources;

[0008] Step 2: Preprocess the collected measurement data containing systematic errors;

[0009] Step 3: Fusion of pre-processed data of the same type;

[0010] Step 4: Convert the fused system error into the ECEF coordinate system to achieve a unified representation of the system error;

[0011] Step 5: Perform Q-type cluster analysis on the measurement data with systematic errors after coordinate transformation from the distance dimension, azimuth dimension and elevation dimension respectively to obtain clustering results with different data characteristics;

[0012] Step 6: Calculate the system errors corresponding to the clustering result measurement data from the three dimensions of distance, azimuth and elevation respectively, collect the data features with the smallest system errors in the corresponding dimensions, and recalculate the system errors corresponding to the data features with the smallest system errors in the three dimensions of distance, azimuth and elevation.

[0013] Preferably, the data source includes a radar system and a ground measuring instrument.

[0014] Preferably, the specific method of preprocessing the collected measurement data containing systematic errors includes: data cleaning, data alignment and correction.

[0015] Preferably, the specific method of converting the fused system error into the ECEF coordinate system is:

[0016] Step 4-1: Convert the measured data from the relative coordinate system to the local Cartesian coordinate system;

[0017] Step 4-2: Transform the target from the local Cartesian coordinate system to the ECEF coordinate system.

[0018] Preferably, the specific formula for converting the measurement data from the relative coordinate system to the local Cartesian coordinate system is:

[0019]

[0020] Among them (r t ,θ t , η t ) is the measured value of the target, (x l ,y l , z l ) is the target local Cartesian coordinate value.

[0021] Preferably, the target is measured from a local Cartesian coordinate system (x l ,y l , z l ) is converted to the ECEF coordinate system (x t ,y t , z t ) is as follows:

[0022]

[0023] In the formula, (x s ,y s , z s ) is the coordinate of the measurement node in ECEF, and R is the rotation matrix.

[0024] Preferably, a Q-type cluster analysis is performed on the measurement data with systematic errors after coordinate transformation from the distance dimension, azimuth dimension and elevation angle dimension, and a specific method for obtaining clustering results with different data features is as follows:

[0025] Step 5-1: Divide the measurement data with systematic errors after coordinate transformation into different categories and clusters;

[0026] Step 5-2: Based on the data characteristics, perform Euclidean similarity calculation, and based on the similarity calculation results, construct a similarity matrix between the measured data, and bring it into the Q-type clustering algorithm to obtain clustering results of different data characteristics.

[0027] Preferably, the specific method of calculating the system error corresponding to the clustering results of different data features from the three dimensions of distance, azimuth and elevation in step 6 is:

[0028] Step 6-1: Construct the equation in the ECEF coordinate system based on the detection results of the two measurement nodes on the same target:

[0029] χ a +R a χ al，k = x b +R b χ bl，k

[0030] in:

[0031]

[0032] χ a and χ b are the ECEF coordinates of the two measurement nodes a and b, R a and R b are the rotation matrices of the two measurement nodes, χ′ al，k and χ′ bl，k are the Cartesian coordinates of the target at the two measurement nodes.

[0033] Step 6-2: Perform a first-order Taylor expansion on the equation in step 6-1:

[0034] χ ae，k +R a J a，k ξ a = x be，k +R b J b，k ξ b

[0035] where χ ae，k and χ be，k are the ECEF coordinates of the target observed by the two nodes, Jae，k and J be，k They are the Jacobian matrices calculated when the system error of the target is 0;

[0036] Step 6-3: Express the formula in 6-2 in matrix form, and substitute it into the ECEF coordinate values ​​of the target observed by the two measurement nodes in the measurement data corresponding to the Q-type clustering result, and solve the system error by the least squares method.

[0037] Preferably, the system error is solved by the least squares method:

[0038] Lξ=Δχ

[0039]

[0040] Where L = [R a J a , R b J b ], ξ = [0, 0, 0, 0, 0] is the initial estimate of the system error without prior information, Δχ is the difference in the ECEF coordinates of the target observed by the two nodes; L is full rank to ensure is uniquely solved to, where Δr a , Δθ a , Δη a , Δr b , Δθ b , Δη b They are the minimum variance unbiased estimates of the two-node system errors in three dimensions: distance, azimuth, and elevation.

[0041] The present invention also proposes a system error correction system based on ECEF and Q-type cluster analysis, comprising:

[0042] Data preprocessing module, used to preprocess the collected measurement data containing systematic errors

[0043] Multi-source data fusion module, used to fuse pre-processed data of the same type;

[0044] ECEF coordinate conversion module, used to convert the fused system error into the ECEF coordinate system to achieve a unified representation of the system error;

[0045] The Q-type cluster analysis module is used to perform Q-type cluster analysis on the measurement data with systematic errors after coordinate transformation from the distance dimension, azimuth dimension and elevation dimension respectively, and obtain clustering results with different data characteristics;

[0046] The system error correction module is used to calculate the system errors corresponding to the clustering result measurement data from the three dimensions of distance, azimuth and elevation, collect the data features with the smallest system errors in the corresponding dimensions, and recalculate the system errors corresponding to the data features with the smallest system errors in the three dimensions of distance, azimuth and elevation.

[0047] Compared with the prior art, the present invention has the following significant advantages: the present invention combines multi-source information fusion, ECEF coordinate system and Q-type cluster analysis to provide a more accurate and efficient system error correction method; by fusing information from multiple data sources, the accuracy and consistency of the data are improved; through ECEF coordinate system conversion and Q-type cluster analysis, potential patterns and laws in the data can be identified, further improving the correction effect of the system error. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of a systematic error correction method based on ECEF and Q-type cluster analysis. DETAILED DESCRIPTION

[0049] A system error correction method based on ECEF and Q-type cluster analysis. This method analyzes and corrects system errors by fusing information from multiple radar data sources, combining the ECEF coordinate system and the Q-type clustering algorithm. It has higher accuracy and consistency than traditional system error correction methods. The processing flow diagram is shown in the figure. Figure 1 As shown, in combination with the embodiment, the specific steps are as follows:

[0050] Step 1: Collect multi-source data: Collect measurement data containing system errors from different data sources, including radar systems, ground measurement instruments, etc., to obtain more comprehensive and diverse data. At the same time, obtain the target AIS, Beidou or GPS real value data, which is only used for comparison of calculated system error results;

[0051] Step 2: Data preprocessing: preprocess the collected measurement data containing systematic errors, including data cleaning, data alignment and correction, to improve the accuracy and consistency of the data;

[0052] Step 3: Multi-source data fusion: fuse the pre-processed data of the same type and integrate the information from different data sources to extract more accurate and reliable system error information;

[0053] Step 4: ECEF coordinate system conversion: Convert the fused system error into the ECEF coordinate system to achieve unified representation and processing of the system error:

[0054] Step 4-1: Filter the measured data and convert them from the relative coordinate system to the local Cartesian coordinate system:

[0055]

[0056] Among them (r t ,θ t , η t ) is the target measurement data, (x l ,y l , z l ) is the local Cartesian coordinate value of the target;

[0057] Step 4-2: Move the target from the local Cartesian coordinate system (x l ,y l , z l ) is converted to the ECEF coordinate system (x t ,y t , z t ):

[0058]

[0059]

[0060] Where (x s ,y s , z s ) is the coordinate of the measurement node (such as radar) in ECEF, and R is the rotation matrix.

[0061] Step 5: Q-type cluster analysis: Perform Q-type cluster analysis on the measurement data with systematic errors after coordinate transformation, and divide the data into different categories and clusters to identify the potential patterns and rules in the data. The accuracy of the systematic errors calculated for data information with different characteristics is different. Through Q-type cluster analysis, observe the accuracy of the systematic errors calculated under different data characteristics, summarize the data characteristics, and collect the characteristic data for systematic error correction, which can further improve the accuracy of systematic error correction:

[0062] Step 5-1: Divide the measurement data with systematic errors after coordinate transformation into different categories and clusters, such as expanding them from three dimensions of distance, azimuth, and elevation, and setting different features;

[0063] Step 5-2: Based on a certain type of data feature, perform Euclidean similarity calculation, and based on the similarity calculation results, construct a similarity matrix between the measured data, and bring it into the Q-type clustering algorithm to obtain clustering results of different data features.

[0064] Step 6: Calculate the systematic errors corresponding to the clustering results of different data features from the three dimensions of distance, azimuth and elevation, collect the data features with the smallest systematic errors in the corresponding dimensions, and recalculate the systematic errors corresponding to the data features with the smallest systematic errors in the three dimensions of distance, azimuth and elevation.

[0065] Step 6-1: Construct the equation in the ECEF coordinate system based on the detection results of the two measurement nodes of the same target:

[0066] χ a +R a χ al，k = x b +R b χ bl，k (4)

[0067] in:

[0068]

[0069] χ a and χ b are the ECEF coordinates of the two measurement nodes a and b, R a and R b are the rotation matrices of the two measurement nodes, χ′ al，k and χ′ bl，k are the Cartesian coordinates of the target at the two measurement nodes.

[0070] Step 6-2: Assuming that the system deviation is small, expand the formula in step 6-1 into the first order Taylor expansion:

[0071] χ ae，k +R a J a，k ξ a = x be，k +R b J b，k ξ b (6)

[0072] where χ ae，k and χ be，k are the ECEF coordinates of the target observed by the two nodes, J ae，k and J be，k They are the Jacobian matrices calculated when the system error of the target is 0;

[0073] Step 6-3: Express the formula in 6-2 in matrix form, substitute the ECEF coordinate values ​​of the target observed by the two measurement nodes, and solve the system error by the least squares method:

[0074]

[0075] Where L = [R a J a , R b J b ], ξ = [0, 0, 0, 0, 0] is the initial estimate of the system error without prior information, Δχ is the difference in the ECEF coordinates of the target observed by the two nodes, and L is a full rank column to ensure can be uniquely solved to, where Δr a , Δθ a , Δη a , Δr b , Δθ b , Δη b They are the minimum variance unbiased estimates of the system errors of the two nodes in the three dimensions of distance, azimuth and elevation respectively;

[0076] Step 6-4: Based on the calculation results of step 6-3, respectively obtain the data with the smallest system error in distance, azimuth, and elevation under the current measurement data, and use the method of step 6-3 to calculate the system error corresponding to the data with the smallest system error in distance, azimuth, and elevation.

[0077] A system error correction system based on ECEF and Q-type cluster analysis, comprising:

[0078] Data preprocessing module, used to preprocess the collected measurement data containing systematic errors

[0079] Multi-source data fusion module, used to fuse pre-processed data of the same type;

[0080] ECEF coordinate conversion module, used to convert the fused system error into the ECEF coordinate system to achieve a unified representation of the system error;

[0081] The Q-type cluster analysis module is used to perform Q-type cluster analysis on the measurement data with systematic errors after coordinate transformation from the distance dimension, azimuth dimension and elevation dimension respectively, and obtain clustering results with different data characteristics;

[0082] The system error correction module is used to calculate the system errors corresponding to the clustering result measurement data from the three dimensions of distance, azimuth and elevation, collect the data features with the smallest system errors in the corresponding dimensions, and recalculate the system errors corresponding to the data features with the smallest system errors in the three dimensions of distance, azimuth and elevation.

Claims

1. A method for correcting system errors based on ECEF and Q-type cluster analysis, characterized in that: include: Step 1: Collect measurement data containing systematic errors from different data sources; Step 2: Preprocess the collected measurement data containing systematic errors; Step 3: Fusion of pre-processed data of the same type; Step 4: Convert the fused system error into the ECEF coordinate system to achieve a unified representation of the system error; Step 5: Perform Q-type cluster analysis on the measurement data with systematic errors after coordinate transformation from the distance dimension, azimuth dimension and elevation dimension respectively to obtain clustering results with different data characteristics; Step 6: Calculate the system errors corresponding to the clustering result measurement data from the three dimensions of distance, azimuth and elevation respectively, collect the data features with the smallest system errors in the corresponding dimensions, and recalculate the system errors corresponding to the data features with the smallest system errors in the three dimensions of distance, azimuth and elevation.

2. The method for correcting system errors based on ECEF and Q-type cluster analysis according to claim 1, characterized in that: The data sources include radar systems and ground measurement instruments.

3. The method for correcting system errors based on ECEF and Q-type cluster analysis according to claim 1, characterized in that: The specific methods for preprocessing the collected measurement data containing systematic errors include: data cleaning, data alignment and correction.

4. The method for correcting system errors based on ECEF and Q-type cluster analysis according to claim 1, characterized in that: The specific method of converting the fused system error into the ECEF coordinate system is: Step 4-1: Convert the measured data from the relative coordinate system to the local Cartesian coordinate system; Step 4-2: Transform the target from the local Cartesian coordinate system to the ECEF coordinate system.

5. The method for correcting system errors based on ECEF and Q-type cluster analysis according to claim 4, characterized in that: The specific formula for converting the measured data from the relative coordinate system to the local Cartesian coordinate system is: Among them (r t ,θ t , η t ) is the measured value of the target, (x l ,y l , z l ) is the target local Cartesian coordinate value.

6. The method for correcting system errors based on ECEF and Q-type cluster analysis according to claim 4, characterized in that: The target is transformed from the local Cartesian coordinate system (x l ,y l , z l ) is converted to the ECEF coordinate system (x t ,y t , z t ) is as follows: In the formula, (x s ,y s , z s ) is the coordinate of the measurement node in ECEF, and R is the rotation matrix.

7. The method for correcting system errors based on ECEF and Q-type cluster analysis according to claim 1, characterized in that: The specific method of performing Q-type cluster analysis on the measurement data with systematic errors after coordinate transformation from the distance dimension, azimuth dimension and elevation angle dimension to obtain clustering results with different data characteristics is as follows: Step 5-1: Divide the measurement data with systematic errors after coordinate transformation into different categories and clusters; Step 5-2: Based on the data characteristics, perform Euclidean similarity calculation, and based on the similarity calculation results, construct a similarity matrix between the measured data, and bring it into the Q-type clustering algorithm to obtain clustering results of different data characteristics.

8. The method for correcting system errors based on ECEF and Q-type cluster analysis according to claim 1, characterized in that: Step 6 The specific method for calculating the system error corresponding to the clustering results of different data features from the three dimensions of distance, azimuth and elevation is: Step 6-1: Construct the equation in the ECEF coordinate system based on the detection results of the two measurement nodes on the same target: x a +R a x' al,k =x b +R b x' bl,k in: χ a and χ b are the ECEF coordinates of the two measurement nodes a and b, R a and R b are the rotation matrices of the two measurement nodes, χ' al,k and χ' bl,k They are the Cartesian coordinates of the target at the two measurement nodes'' Step 6-2: Perform a first-order Taylor expansion on the equation in step 6-1: x ae,k +R a J a,k x a =x be,k +R b J b,k x b where χ ae，k and χ be，k are the ECEF coordinates of the target observed by the two nodes, J ae，k and J be，k They are the Jacobian matrices calculated when the system error of the target is 0; Step 6-3: Express the formula in 6-2 in matrix form, and substitute it into the ECEF coordinate values ​​of the target observed by the two measurement nodes in the measurement data corresponding to the Q-type clustering result, and solve the system error by the least squares method.

9. The method for correcting system errors based on ECEF and Q-type cluster analysis according to claim 8, characterized in that: Solve the system error by the least squares method: Lξ=Δχ Where L = [R a J a , R b J b ], ξ=[0,0,0,0,0,0] is the initial estimate of the system error without prior information, Δχ is the difference in the ECEF coordinates of the target observed by the two nodes; L is full rank to ensure is uniquely solved to, where Δr a ,Δθ a ,Δη a ,Δr b ,Δθ b ,Δη b They are the minimum variance unbiased estimates of the two-node system errors in three dimensions: distance, azimuth, and elevation.

10. A system error correction system based on ECEF and Q-type cluster analysis using the method of any one of claims 1 to 9, characterized in that: include: Data preprocessing module, used to preprocess the collected measurement data containing systematic errors Multi-source data fusion module, used to fuse pre-processed data of the same type; ECEF coordinate conversion module, used to convert the fused system error into the ECEF coordinate system to achieve a unified representation of the system error; The Q-type cluster analysis module is used to perform Q-type cluster analysis on the measurement data with systematic errors after coordinate transformation from the distance dimension, azimuth dimension and elevation dimension respectively, and obtain clustering results with different data characteristics; The system error correction module is used to calculate the system errors corresponding to the clustering result measurement data from the three dimensions of distance, azimuth and elevation, collect the data features with the smallest system errors in the corresponding dimensions, and recalculate the system errors corresponding to the data features with the smallest system errors in the three dimensions of distance, azimuth and elevation.

Images