Dynamic rapid UWB base station self-calibration method implemented on label

Abstract

Aiming at the current situation that a UWB base station needs to be manually arranged for measurement and manually burnt to a to-be-used label at present, time and labor are wasted, and errors are large, the invention provides a dynamic and rapid UWB base station self-calibration method implemented on the label. According to the method, a single chip microcomputer and other controllers are used for sending a signal to a label to be used, then bilateral distance measurement is carried out between each base station and a note, dimension reduction processing is carried out through measurement MDS, truncated singular value decomposition is carried out on an inner product matrix, the process is repeated to obtain multiple results to estimate a central value, impurities are filtered out after a threshold value is set according to needs, and the central value is calculated again. And finally, high-accuracy base station coordinates are obtained by translating the coordinate system. The addition and subtraction base stations only need to send signals, manual intervention is not needed in the whole process, and the self-calibration method is rapid, convenient and accurate.

Description

Technical Field

[0001] This invention belongs to the field of autonomous positioning, and in particular relates to a dynamic and fast UWB base station self-calibration method implemented on a tag. Background Technology

[0002] Ultra-wideband (UWB) technology is a popular wireless carrier communication technology that uses nanosecond-level non-sinusoidal narrow pulses to transmit data. UWB technology has advantages such as low system complexity, low transmitted signal power spectral density, insensitivity to channel fading, low interception capability, and high positioning accuracy, making it widely used in positioning in complex indoor and outdoor scenarios. UWB uses the Two-Way Ranging (TWR) method for positioning, calculating the distance between the tag and the base station using the round-trip time of flight (TOF), and finally calculating the tag's position using a trilateration algorithm. This method requires both the tag and the base station to have the ability to transmit and receive signals simultaneously. During measurement, both signal transmission and reception are required, and the distance between the tag and the base station is measured by calculating the time intervals between their respective transmission and reception. This method requires the deployment of three or more base stations to form a coordinate system for UWB tag positioning, and its positioning accuracy is significantly affected by the calibration coordinates of the base stations themselves.

[0003] There are two main types of existing base station calibration technologies. One type involves manually measuring the coordinates of each base station after deployment and then writing them into the corresponding program on the tag. The other type enables communication between base stations, using a base station as the in-place coordinate to calculate the positions of the remaining base stations via TWR. However, the coordinates of the UWB base stations still need to be manually written into the tag. Summary of the Invention

[0004] To overcome the above-mentioned shortcomings, this invention provides a dynamic and fast UWB base station self-calibration method implemented on a tag, which improves the self-calibration accuracy and reduces the self-calibration steps, making the process simpler.

[0005] This invention is implemented using the following technical solution, including the following steps:

[0006] Step 1: Deploy UWB base stations and tags. This requires three or more base stations and one tag within the range of each base station.

[0007] Step 2: Control the UWB tags using a microcontroller or other controller to send a self-calibration start signal. Each UWB base station and tag communicates with each other, and the base station's identifier can be extracted. The communication results are stored in the tag for later parsing.

[0008] Step 3: Analyze the distances between points to form a distance matrix, and then use the metric MDS method. The inner product matrix between points is calculated using the distance matrix.

[0009] Step 4: Perform truncated singular value decomposition on the inner product matrix. The goal is to obtain two-dimensional coordinates, and finally, the two corresponding eigenvalues ​​and the corresponding eigenvector matrix.

[0010] Step 5: Repeat the above steps multiple times to perform the mean-based outlier removal method. Calculate the average of the coordinates, filter out noise based on the set threshold, and then recalculate the average.

[0011] Step Six: Scale the coordinates, setting base station 0 as the zero point in the XOY coordinate system. Distribute the remaining base stations and tags as much as possible in the upper half of the coordinate system. The tags directly obtain the positions of each base station and their own initial positions, and send a self-calibration completion signal to the microcontroller or controller.

[0012] Step one includes: deploying three or more base stations, with a maximum capacity of eight base stations, and the number of base stations can be dynamically increased or decreased at any time.

[0013] Step two includes: a microcontroller sending a specific self-calibration signal to the UWB tag, which instructs each base station and the tag to perform TWR communication. The TWR method calculates the distance between points by measuring the Time-of-Flight (ToF) of the UWB radio frequency signal and then multiplying it by the speed of light. The resulting raw ranging information is then stored in the UWB tag for further processing.

[0014] Step three includes: forming a distance matrix based on the distances between points. For ease of calculation, the distances are treated as scalars; for example, the distance from base station 1 to base station 2 is equal to the distance from base station 2 to base station 1. The resulting distance matrix is ​​a symmetric matrix. Then, the inner product matrix is ​​obtained by equating the high-dimensional distance with the low-dimensional Euclidean distance. The dimensionality-reduced sample coordinate matrix needs to be centered. Centering involves subtracting the mean vector of the entire sample set from each sample vector. Therefore, summing all sample vectors yields a zero vector.

[0015] Step four includes: performing truncated singular value decomposition on the inner product matrix. Due to the dimensionality reduction requirement of this algorithm, we only retain the eigenvalues ​​of the first two columns and form a 5*2 eigenvector matrix corresponding to the eigenvalues, which greatly speeds up the computation efficiency.

[0016] Step five includes: calculating the average of the horizontal and vertical coordinates for, for example, 50 coordinates; analyzing the distance of each point from the average value; excluding points whose distance is greater than 1.5 times the average value; recalculating the average value for the remaining samples; and using this point as the final confirmed coordinate.

[0017] Step six includes: after the above calculation, the zero point is near the tag, calculate the difference between the coordinates of base station 0 and the coordinates of the zero point, and move the entire coordinate system toward base station 0. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the main logic implementation of the solution in a specific example provided in the embodiments of the present invention;

[0020] Figure 2 This is a schematic diagram of the UWB base station and tag layout and communication provided in an embodiment of the present invention; Detailed Implementation

[0021] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are merely descriptive and not limiting, and should not be used to limit the scope of protection of the present invention.

[0022] Figure 1 This is a flowchart illustrating the overall implementation of the present invention.

[0023] Step 1: Deploy base stations and tag locations. The self-calibration signal can be transmitted via serial port, Bluetooth, WiFi, etc. Before starting ranging, ensure that at least three base stations are deployed, and the tags are placed anywhere within the area enclosed by the base stations.

[0024] Step 2: TWR communication is established between the base station and the tag, using T... prop Indicates flight time:

[0025]

[0026] The distance between points is obtained by multiplying the distance by the speed of light through communication and stored in a tag.

[0027] Step 3: After forming the distance matrix, reduce the dimension to form the relevant inner product matrix, as follows:

[0028] (1) Form a distance matrix D, taking 4 base stations (A0, A1, A2 and A3) and 1 tag (T) as an example:

[0029]

[0030] The default distance between each point and itself is 0, and it is stipulated that:

[0031] distij =dist ji (3)

[0032] (2) Calculate the inner product matrix using the distance matrix. First, calculate the centralized matrix H, where n represents the total number of base stations and tags:

[0033]

[0034] (3) Calculate the inner product matrix A, D×D T The matrix representing the squared distance:

[0035]

[0036] Step 4: Perform truncated singular value decomposition on A, where Λ = diag(λ1, λ2, ..., λ). d Let be a diagonal matrix composed of eigenvalues, where λ1≥λ2≥…≥λ d V represents the corresponding eigenvector matrix:

[0037] A = VΛV T (6)

[0038] Since we intend to reduce the dimensionality to 2-dimensional space and only retain the x-axis and y-axis coordinates of each point, we retain 2 eigenvalues. The final V is a 5*2 matrix and Λ is a 2*2 matrix.

[0039] Step 5: Process the coordinates obtained from each point as follows:

[0040] (1) Assuming that m experiments were conducted, we first need to calculate the average values ​​of the x-axis and y-axis for each point:

[0041]

[0042]

[0043] (2) Calculate the standard deviation for each point:

[0044]

[0045]

[0046] (3) Obtain the coordinate threshold like or Then it is considered an error point and eliminated.

[0047] (4) Calculate the average of the coordinates of the k points (k≤m) within the threshold again to obtain the coordinates of each point.

[0048] Step 6: Translate the coordinates of base station 0 to the zero point (0, 0), and calculate the required translation length:

[0049]

[0050]

[0051] Finally, the base station coordinates and tag coordinates can be obtained, and subsequent work can begin directly.

Claims

1. A dynamic fast UWB base station self-calibration method implemented on a tag, characterized by: Step one: UWB base station and tag arrangement, 3 or more base stations are needed, and 1 tag within the base station range. Step two: control the UWB tag through a single-chip microcomputer or other controller, send a start signal, and communicate multiple times between each UWB base station and the single tag, while extracting the base station number. Store the communication results on the tag. Step three: form a distance matrix between each point, the distance between each point in the UWB communication process is the Euclidean distance, so the metric MDS method is used below. Convert the distance matrix between each point to an inner product matrix. Step four: perform truncated singular value decomposition on the inner product matrix to obtain two-dimensional coordinates. According to the method requirements, obtain the corresponding two eigenvalues and eigenvector matrix, and the calculation is very fast. Step five: repeat the above work multiple times, estimate the center point, set a threshold to filter out error points, and then calculate the average value again. This value is the coordinate of each point. Step six: scale the coordinates, set base station 0 as the zero point in the XOY coordinate system, and distribute the remaining base stations and tags in the upper half of the coordinate system. The tag directly obtains the positions of each base station and its initial position, and sends a self-calibration signal to the single-chip microcomputer or controller.

2. The method of claim 1, wherein, In step two, the base station and the tag communicate with each other, and the communication results are stored in the tag for calculation, without occupying the base station or single-chip microcomputer calculation space.

3. The method of claim 1, wherein, In step three, a distance matrix is formed between each point. For convenience of calculation, the distance between base stations is equal, so the distance matrix formed is a symmetric matrix. The inner product matrix is obtained by equating high-dimensional distance to low-dimensional Euclidean distance. The dimension-reduced sample coordinate matrix needs to be centralized.

4. The method of claim 1, wherein, In step four, perform truncated singular value decomposition on the inner product matrix. Regardless of the number of base stations, we only need to obtain the first two eigenvalues and the corresponding n*2 eigenvector matrix, and the calculation is fast.

5. The method of claim 1, wherein, In step five, the horizontal and vertical coordinates of multiple coordinates need to be calculated, and the average value of each point distance is analyzed. If the distance is greater than 1.5 times the average value, it is considered an error point and excluded. The remaining samples are recalculated to obtain the final confirmed coordinate.

6. The method of claim 1, wherein, In step six, the coordinate system is translated, and base station 0 is set as the origin. The tag obtains the positions of each base station and its initial position, and directly starts working.

7. The method of claim 4, wherein, First, obtain the symmetric distance matrix D, then calculate the centering matrix H, and finally obtain the inner product matrix A. The default distance between each point and itself is 0 dist ij = dist ji (2) 8. The method of claim 5, wherein, Perform truncated singular value decomposition on A, Λ = diag(λ1, λ2) is a diagonal matrix composed of eigenvalues, λ1 ≥ λ2. V represents the corresponding eigenvector matrix: A = VΛV T (5).

9. The method of claim 6, wherein, Process the coordinates of each point. Assuming m experiments are performed, we first need to calculate the average value and standard deviation of the x-axis and y-axis of each point: Obtaining coordinate threshold value And If Or It is regarded as an error point and is eliminated. The coordinates of the k points (k≤m) within the threshold value are averaged again, and finally the coordinates of each point are obtained.

10. The method of claim 7, wherein, Translate the coordinates of base station 0 to the coordinate zero point (0, 0), calculate the length of the translation, and then translate the coordinates of each point: Finally, the base station coordinates and tag coordinates can be obtained, and the subsequent work can be directly started.

Images