A positioning accuracy estimation method based on KNN

Abstract

The present invention provides a method for estimating positioning accuracy based on KNN, belonging to the technical field of navigation and positioning. It includes three parts: an offline data collection stage, an online data collection and an online estimation stage. The specific steps are as follows: First, collect data at different reference points to form a database; Second, collect the vectors of the test points; Third, use the KNN algorithm to achieve error estimation. The advantage of the present invention is that when estimating the positioning accuracy, it not only takes into account the influence of the time difference of arrival between the terminal and the base station on the error, but also takes into account the influence of the spatial distribution between the terminal and the base station on the error, which can improve the accuracy of the positioning error estimation.

Description

Technical Field

[0001] The present invention belongs to the technical field of navigation and positioning, involves judging the accuracy of positioning, and particularly involves estimating the error of Observed Time Difference of Arrival (OTDOA). Background Art

[0002] With the rapid development of information technology, positioning services have become a popular research direction at present, and people's production and life are inseparable from positioning services. In order to evaluate the accuracy of the results of each positioning, a positioning accuracy estimation algorithm is required. After retrieving existing literature, it is found that "Fusion Localization Based on Accuracy Estimation" published in the International Conference on Advanced Electronic Materials, Computers and Software Engineering (AEMCSE) (the 2022 International Conference on Advanced Electronic Materials, Computers and Software Engineering, "Fusion Localization Based on Accuracy Estimation") proposes a positioning accuracy estimation algorithm. This algorithm only considers using the monotonic characteristics presented by the "set of time differences of arrival between the terminal and the base station" and the "positioning accuracy" to achieve accuracy estimation.

[0003] The positioning accuracy estimation algorithm based on the K-Nearest Neighbors (KNN) position fingerprint algorithm in the present invention not only utilizes the monotonic relationship between the "set of time differences of arrival between the terminal and the base station" and the "positioning accuracy", but also takes into account the influence of the distance and angle between the positioning base station and the terminal device on the positioning result. Compared with previous studies, the advantage of the present invention is that more parameters are involved in the positioning accuracy estimation, and the positioning error is estimated more accurately. Summary of the Invention

[0004] The technical problem to be solved by the present invention is that the original accuracy estimation method only considers the relationship between the "set of time differences of arrival" and the "positioning accuracy", resulting in low accuracy of positioning accuracy estimation. The purpose of the present invention is to improve on the basis of the original accuracy estimation method, increase the parameters for error estimation, and improve the accuracy of positioning accuracy estimation.

[0005] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0006] A positioning accuracy estimation method based on KNN includes three parts: an offline acquisition stage, an online acquisition and an online estimation stage. The specific steps are as follows:

[0007] Step 1: Collect data of different reference points to form a database.

[0008] Step 1.1: Assume there are M base stations participating in OTDOA positioning in the positioning environment. Select the base station with the maximum transmission power as the master base station, and the remaining base stations as slave base stations. Determine the position of the selected reference point (AP), and calculate the average value of the time difference set of the terminal device at the reference point (AP). where τ i corresponds to the time when the i-th base station reaches the terminal device.

[0009] Step 1.2: Calculate the average value of the included angles between the terminal device at the reference point (AP) and the master base station and slave base stations. where θ i is the included angle between the two straight lines formed by the i-th base station to the terminal device and the master base station to the terminal device, and is obtained by calculating the position coordinates among the terminal device, the master base station and each slave base station. This step takes into account the influence of the spatial angle distribution between the terminal and the base stations on the error.

[0010] Step 1.3: Calculate the average value of the distances between the terminal device at the reference point (AP) and each base station in the two-dimensional plane. where, R i is the distance between base station i and the terminal device. This step takes into account the influence of the spatial distance distribution between the terminal and the base stations on the error.

[0011] Step 1.4: Form a reference vector X s =(C, θ e , R e ) at the reference point (AP), and calculate the positioning error E s of the terminal device at this reference point (AP). Store the reference vector X s and the positioning error E s to form a database.

[0012] Step 1.5: Evenly distribute W reference points (AP) in the area. Repeat steps 1.1 to 1.4 at each reference point (AP), and repeat N times. Store W*N reference vectors X s and the corresponding error E s to obtain a database.

[0013] Step 2: Collect the vectors of the test points.

[0014] Step 2.1: Calculate the average value of the time difference set of the terminal device at the test point (TP). where τ i corresponds to the time when the i-th base station reaches the terminal device.

[0015] Step 2.2: Calculate the average value of the angles between the terminal device and the master base station and the slave base station in the two-dimensional plane when the terminal device is at the test point (TP) according to the positioning result and the position coordinates of the base stations.

[0016] Step 2.3: Calculate the average value of the distances between the terminal device and each base station in the two-dimensional plane when the terminal device is at the test point (TP) according to the positioning result and the position coordinates of the base stations.

[0017] Step 2.4: Form the received vector X of the terminal device at the test point (TP). t =(C t , θ t , R t ).

[0018] The third step is to use the KNN algorithm to achieve error estimation.

[0019] Step 3.1: Calculate the Euclidean distances between the received vector X t in Step 2.4 and multiple reference vectors X s in the database respectively.

[0020] Step 3.2: Arrange the obtained Euclidean distances in ascending order. According to the KNN algorithm, take the error information E s corresponding to the first k reference points (where k ≤ W), and calculate the mean value, which is the error estimation value of the current positioning.

[0021] The effects and benefits of the present invention are as follows: Compared with previous studies, the advantage of the present invention is that when estimating the positioning accuracy, it not only considers the influence of the time difference of arrival between the terminal and the base stations on the error, but also considers the influence of the spatial distribution between the terminal and the base stations on the error, thereby improving the accuracy of the positioning error estimation. Description of the Drawings

[0022] Figure 1 It is a schematic diagram of Step 1.2 in the content of the invention. Taking 4 base stations participating in the positioning as an example, θ2, θ3, and θ4 respectively correspond to the angles between base stations 2, 3, 4, the master base station, and the terminal device. Detailed Embodiments

[0023] The following further describes the present invention with specific embodiments.

[0024] The first step is to collect data at different reference points to form a database.

[0025] Step 1.1: Assume that there are 5 base stations participating in OTDOA positioning in the positioning environment, including one macro base station and four micro base stations. Select the macro base station as the main base station. Determine the position of the reference point (AP), obtain the time-of-arrival τ1, …, τ5 between each base station and the terminal, and calculate the average value of the set of time differences of the terminal device at the reference point (AP).

[0026] Step 1.2: Taking the terminal device as the vertex, calculate the angles θ2, …, θ5 between the terminal device, the main base station and each slave base station at the reference point (AP) in the two-dimensional plane, and take the average value of the angles.

[0027] Step 1.3: Calculate the distances R1, …, R5 between the terminal device and each base station at the reference point (AP) in the two-dimensional plane, and take the average value of the distances to obtain

[0028] Step 1.4: Form the reference vector X at the reference point (AP). s =(C, θ e , R e ), and statistically calculate the positioning error E of the terminal device at this reference point (AP). s , and store the reference vector X s and the positioning error E s to form a database.

[0029] Step 1.5: Evenly distribute 10 5 reference points (AP) in the area. Repeat Steps 1.1 to 1.4 at each reference point (AP), and repeat 10 times. Store 10 6 reference vectors X s and the corresponding errors E s to obtain a database

[0030] Second step, collect the vectors of the test points.

[0031] Step 2.1: Calculate the average value of the set of time difference of arrival at the test point (TP) of the terminal device. where τ1, …, τ5 are 3.334×10 -7 , 1.6665×10 -7 , 5.002×10 -7 , 1.02×10 -6 , 3.32×10 -7 .

[0032] Step 2.2: Calculate the average value of the angles between the terminal device and each base station in the two-dimensional plane at the test point (TP). where θ2, ..., θ5 are 29.8°, 75.6°, 45.2°, 108.5° respectively.

[0033] Step 2.3: Calculate the average distance between the terminal device and each base station on the two-dimensional plane under the condition at the test point (TP). where R1, ..., R5 are 105.2 meters, 48.9 meters, 158.7 meters, 310.5 meters, 107.8 meters respectively.

[0034] Step 2.4: Form the received vector X of the terminal device at the test point (TP). t =(C t , θ t , R t ).

[0035] Thirdly, use the KNN algorithm to achieve error estimation.

[0036] Step 3.1: Calculate the received vector X in Step 2.4 t and the Euclidean distance between the reference vector X s in the database.

[0037] Step 3.2: Arrange the obtained Euclidean distances in ascending order. According to the KNN algorithm, take the error information E s corresponding to the first 7 reference points, and obtain the average value, which is the error estimation value of the current positioning.

[0038]

[0039] According to the result obtained in Step 3.2, it can be seen that in this positioning, the estimation deviation obtained by using the present invention is 10.2 meters, and the estimation deviation obtained by the traditional method is 15.6 meters. The estimation error method proposed by the present invention is more accurate.

[0040] The above embodiments only represent the implementation manners of the present invention, but should not be construed as limiting the scope of the present invention patent. It should be noted that for those skilled in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. A positioning accuracy estimation method based on KNN, characterized in that The method described above includes three parts: an offline acquisition stage, an online acquisition and an online estimation stage, and comprises the following steps: First, collect data at different reference points to form a database; Step 1.1: Assume that there are M base stations participating in OTDOA positioning in the positioning environment. Select the base station with the maximum transmission power as the master base station, and the remaining base stations as slave base stations; determine the position of the reference point (AP), and calculate the average value of the time difference set of the terminal device at the reference point AP where τ i corresponds to the time when the i-th base station reaches the terminal device; Step 1.2: Calculate the average value of the angles between the terminal device and the master base station and the slave base stations at the reference point AP where θ i is the angle between the two straight lines formed by the i-th base station to the terminal device and the master base station to the terminal device, and is obtained by calculating the position coordinates among the terminal device, the master base station, and each slave base station; Step 1.3: Calculate the average value of the distances between the terminal device at the reference point AP and each base station in the two-dimensional plane where R i is the distance between base station i and the terminal device; Step 1.4: Form the reference vector X at the reference point AP s =(C, θ e , R e ), and calculate the positioning error E of the terminal device at this reference point AP s , store the reference vector X s and the positioning error E s for preservation to form a database; Step 1.5: Evenly distribute W reference points AP within the area. Repeat Steps 1.1 to 1.4 at each reference point AP and repeat N times, storing W * N reference vectors X s and the corresponding error E s , to obtain a database; Second, collect the vectors of the test points; Step 2.1: Calculate the average value of the set of time difference of arrival at the test point TP of the terminal device where τ i corresponds to the time when the i-th base station reaches the terminal device; Step 2.2: Calculate the average value of the angles between the terminal device and the master base station and the slave base station on the two-dimensional plane when the terminal device is at the test point TP according to the positioning result and the position coordinates of the base stations Step 2.3: Calculate the average distance between the terminal device and each base station on the two-dimensional plane when the terminal device is at the test point TP according to the positioning result and the position coordinates of the base stations Step 2.4: Form the received vector X of the terminal device at the test point TP t =(C t , θ t , R t ); Third, use the KNN algorithm to implement error estimation; Step 3.1: Calculate the received vector X in Step 2.4 t and the multiple reference vectors X s in the database respectively, and calculate the Euclidean distance between them; Step 3.2: Arrange the obtained Euclidean distances in ascending order. According to the KNN algorithm, take the error information E corresponding to the first k reference points s , and calculate the mean value, which is the error estimation value of the current positioning where k ≤ W.

Images