A WiFi fingerprint positioning method based on minimum coverage circle of reference points

Abstract

The present invention discloses a WiFi fingerprint positioning method based on the minimum coverage circle of reference points. The method uses the minimum coverage circle of reference points to determine whether there are reference points that are far apart but have similar position fingerprints. When the abnormal phenomenon occurs, the selected position fingerprint is filtered by the position of the person being located at the previous moment to reduce the influence of reference points that are far apart but have similar position fingerprints on the positioning accuracy. In addition, the present invention also designs a fusion distance mechanism and a position fingerprint automatic selection mechanism. It has been experimentally confirmed that the present invention has higher accuracy than other traditional positioning algorithms.

Description

Technical Field

[0001] The present invention relates to the field of wireless sensor network positioning, and in particular to an indoor positioning algorithm using RSS of WiFi signals. Background Art

[0002] With the construction of smart cities, indoor positioning technology has become increasingly important. It plays a vital role in smart warehousing, precision marketing, emergency security, etc. Among the many indoor positioning technologies, WiFi positioning technology has become one of the most promising positioning methods in the field of indoor positioning due to its advantages such as low deployment cost and high accuracy.

[0003] WiFi positioning can be divided into positioning based on ranging and positioning based on location fingerprinting according to the principle. Due to the non-line-of-sight effect and multipath effect caused by walls, furniture, pedestrians, etc. in indoor environments, it is difficult to achieve high-precision ranging, so the positioning method based on ranging has low accuracy.

[0004] The location fingerprint-based positioning method has good robustness to non-line-of-sight effects and multipath effects, and is the mainstream method in the field of WiFi positioning. For example, in Chinese patent literature, the invention patent with publication number CN115442740A provides a location fingerprint positioning method and positioning system, which uses a high-sensitivity, low-power LPWAN positioning base station to realize wireless signal fingerprint collection and communication, which can reduce the number of positioning base station deployments; the combination of TOA ranging and fingerprint matching can improve positioning accuracy.

[0005] The location fingerprint-based positioning method can be divided into the RSS fingerprint-based positioning method and the CSI fingerprint-based positioning method according to the different location fingerprints. Although the CSI fingerprint-based positioning method has higher accuracy, most smartphones do not have the ability to collect CSI fingerprints. Therefore, the RSS fingerprint-based positioning method has better universality.

[0006] Ideally, the location fingerprints of the same location are identical, and the location fingerprints of different locations are different. However, in indoor environments with interference factors such as non-line-of-sight effects and multipath effects, the ideal signal attenuation model is no longer applicable, and the signal strength is no longer simply inversely proportional to the signal propagation distance. The location fingerprint is randomly disturbed, and its distribution law shows a certain degree of randomness. Similar location fingerprints may appear at locations far apart. This abnormal phenomenon is misleading to the positioning algorithm and will increase the positioning error. Summary of the invention

[0007] The purpose of the present invention is to provide a new WiFi fingerprint positioning method based on a minimum coverage circle of reference points to overcome some technical problems still existing in the existing fingerprint positioning methods.

[0008] Based on the first aspect of the present invention, a WiFi fingerprint positioning method based on the minimum coverage circle of reference points is proposed. The method uses the minimum coverage circle of reference points to determine whether there are reference points that are far away but have similar location fingerprints. When this abnormal phenomenon occurs, the selected location fingerprint is filtered by the location of the person being located at the previous moment to reduce the influence of reference points that are far away but have similar location fingerprints on the positioning accuracy. This method solves the problem that reference points that are far away but have similar location fingerprints affect the positioning accuracy.

[0009] Specifically, the WiFi fingerprint positioning method based on the minimum coverage circle of the reference point includes at least the following steps:

[0010] S1, calculate the fusion distance between the location fingerprint uploaded by the locatee and all the location fingerprints in the location fingerprint database;

[0011] S2, arrange the fusion distances in ascending order and calculate the difference between adjacent fusion distances;

[0012] S3, find the index value K of the maximum difference;

[0013] S4, rearrange the location fingerprints in the location fingerprint database according to the principle of fusion distance from small to large, and select the first K location fingerprints;

[0014] S5, query the reference points corresponding to the K position fingerprints, calculate the area S of the minimum coverage circle, and set the threshold λ;

[0015] S6, if S<λ, weighted positioning is performed based on the K position fingerprints; if S≥λ, the position fingerprints are filtered in combination with the position of the person being positioned at the previous moment, and then weighted positioning is performed.

[0016] In step S1, the fusion distance between the location fingerprint uploaded by the locatee and all the location fingerprints in the location fingerprint database is calculated as follows:

[0017] FD(FP i ,FP j )=ωED(FP i ,FP j )+(1-ω)CD(FP i ,FP j )

[0018] Among them, FP i and FP j are two different location fingerprints, ED(FP i ,FP j ) is their Euclidean distance, CD(FP i ,FP j) is their cosine distance, ω is the weight of the Euclidean distance, 0<ω<1, FD(FP i ,FP j ) is their fusion distance.

[0019] Furthermore, in step S2, all fusion distances are arranged in ascending order as follows:

[0020] (FD1,FD2,...,FD n )

[0021] Where n is the number of location fingerprints in the location fingerprint database;

[0022] And the difference between two adjacent fusion distances is calculated as follows:

[0023] dif i,j =FD j -FD i

[0024] (dif 1,2 ,dif 2,3 ,...,dif n- 1 ,n ).

[0025] Furthermore, in step S3, the index value of the maximum difference is found as follows:

[0026] K=argmax(dif 1,2 ,dif 2,3 ,...,dif n-1,n )

[0027] And, the index value starts from 1.

[0028] Furthermore, in step S5, the threshold λ is calculated as follows:

[0029] λ=πl 2

[0030] Where l is the minimum spacing between reference points.

[0031] Furthermore, in step S6, if S<λ, weighted positioning is performed according to the following formula based on the K location fingerprints selected from the location fingerprint database:

[0032]

[0033] Among them, FP u is the location fingerprint uploaded by the located person, FPi is the location fingerprint in the i-th selected location fingerprint database, (x i ,y i ) is the location where it is collected, (xu ,y u ) is the predicted position of the located person;

[0034] If S ≥ λ, the selected location fingerprint is filtered in combination with the location of the person being located at the last moment to delete the location fingerprint whose collection location is far from the location of the person being located; the filtering condition is:

[0035] ED(P pre ,P i )>μ

[0036] Among them, P pre is the position of the person being located predicted in the last positioning, P i is the collection location of the i-th location fingerprint, and μ is the maximum walking distance of the person being located within the positioning interval; the location fingerprints that meet this condition are deleted, and weighted positioning is performed according to the following formula based on the filtered K′ location fingerprints:

[0037]

[0038] Among them, FP u is the location fingerprint uploaded by the located person, FPi is the location fingerprint in the i-th selected location fingerprint database, (x i ,y i ) is the location where it is collected, (x u ,y u ) is the predicted position of the person being located.

[0039] Based on the second aspect of the present invention, a WiFi fingerprint positioning system based on a minimum coverage circle of reference points is provided, the system at least comprising:

[0040] Access point, used to send WiFi signals;

[0041] Location fingerprint collection device, a mobile communication device used to collect the RSS of the access point WiFi signal;

[0042] The positioning server stores the location fingerprint database constructed in the offline phase, has a built-in program for executing the aforementioned WiFi fingerprint positioning method based on the minimum coverage circle of reference points, receives the location fingerprint sent by the person being positioned and predicts his or her location.

[0043] Based on the third aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed, the aforementioned WiFi fingerprint positioning method based on the minimum coverage circle of reference points is implemented.

[0044] Based on the fourth aspect of the present invention, a hardware device for WiFi fingerprint positioning includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the aforementioned WiFi fingerprint positioning method based on the minimum coverage circle of reference points is implemented.

[0045] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0046] Firstly, the present invention uses a fusion distance composed of Euclidean distance and cosine distance to measure the similarity of location fingerprints, and comprehensively considers the similarity of location fingerprints in geometric distance and data distribution.

[0047] Secondly, the present invention designs an adaptive K value, dynamically and automatically sets the K value based on the maximum difference of the fusion distance, thereby improving the flexibility and adaptability of the positioning algorithm.

[0048] Finally, the present invention detects whether the positioning is affected by reference points that are far away but have similar position fingerprints through the area size of the minimum coverage circle of the reference points, and when this abnormal phenomenon occurs, the position fingerprint of the collection position that is far away from the position of the person being located at the previous moment is deleted with the help of the position of the person being located, thereby reducing the impact of the abnormal phenomenon on the positioning accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying creative labor, other drawings obtained based on these drawings still belong to the scope of the present invention.

[0050] Figure 1 A flowchart of a WiFi fingerprint positioning method based on a minimum coverage circle of reference points according to an embodiment of the present invention is shown;

[0051] Figure 2 shows the cumulative distribution function of positioning errors of different positioning algorithms according to an embodiment of the present invention;

[0052] Figure 3 A box plot of positioning errors of different positioning algorithms according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0053] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings so that the purpose, features and advantages of the present invention can be more clearly understood. It should be understood that the embodiments shown in the accompanying drawings are not intended to limit the scope of the present invention, but are only intended to illustrate the essential spirit of the technical solution of the present invention.

[0054] In the following description, certain specific details are set forth for the purpose of illustrating the various disclosed embodiments to provide a thorough understanding of the various disclosed embodiments. However, those skilled in the relevant art will recognize that the embodiments may be practiced without one or more of these specific details. In other cases, well-known devices, structures, and techniques associated with the present application may not be shown or described in detail to avoid unnecessarily obscuring the description of the embodiments.

[0055] References throughout the specification to "one embodiment" or "an embodiment" indicate that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of "in one embodiment" or "in an embodiment" in various places throughout the specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any manner in one or more embodiments.

[0056] Embodiment 1:

[0057] In the offline stage, reference points are divided at equal intervals in the area to be positioned, and multiple location fingerprints are collected at the reference points to build a location fingerprint database.

[0058] In the online stage, the person being located sends a location fingerprint to the positioning server. After receiving the location fingerprint of the person being located, the positioning server provides positioning services.

[0059] Calculate the fusion distance between the location fingerprint uploaded by the locatee and all the location fingerprints in the location fingerprint database.

[0060] FD(FP i ,FP j )=ωED(FP i ,FP j )+(1-ω)CD(FP i ,FP j )

[0061] FP i and FP j are two different location fingerprints, ED(FP i ,FP j ) is their Euclidean distance, CD(FP i ,FP j ) is their cosine distance, ω is the weight of the Euclidean distance, 0<ω<1, FD(FP i ,FP j ) is their fusion distance.

[0062] Sort all blend distances in ascending order.

[0063] (FD1,FD2,...,FD n )

[0064] n is the number of location fingerprints in the location fingerprint database.

[0065] Calculate the difference between two adjacent fusion distances.

[0066] dif i,j =FD j -FD i

[0067] (dif 1,2 ,dif 2,3 ,...,dif n-1,n )

[0068] Find the index of the maximum difference. In this algorithm, the index starts at 1.

[0069] K=argmax(dif 1,2 ,dif 2,3 ,...,dif n-1,n )

[0070] The location fingerprints in the location fingerprint database are rearranged according to the principle of fusion distance from small to large, and the first K location fingerprints are selected.

[0071] Find the reference points corresponding to the K position fingerprints, calculate the area S of the minimum coverage circle, and set the threshold λ. The threshold λ needs to be set according to the minimum spacing of the reference points. The calculation method of the threshold λ is as follows:

[0072] λ=πl 2

[0073] l is the minimum spacing of reference points.

[0074] If S<λ, weighted positioning is performed based on K location fingerprints selected from the location fingerprint database.

[0075]

[0076] FP u is the location fingerprint uploaded by the located person, FPi is the location fingerprint in the i-th selected location fingerprint database, (x i ,y i ) is the location where it is collected, (x u ,y u ) is the predicted position of the person being located.

[0077] If S ≥ λ, the selected location fingerprints need to be filtered in combination with the location of the person being located at the last moment, and the location fingerprints whose collection locations are far from the location of the person being located are deleted. The filtering conditions are:

[0078] ED(P pre ,P i )>μ

[0079] P pre is the position of the person being located predicted in the last positioning, P i is the location corresponding to the i-th location fingerprint, and μ is the maximum walking distance of the person being located within the positioning interval. The location fingerprints that meet this condition are deleted, and weighted positioning is performed based on the filtered K′ location fingerprints.

[0080]

[0081] FP u is the location fingerprint uploaded by the located person, FPi is the location fingerprint in the i-th selected location fingerprint database, (x i ,y i ) is the location where it is collected, (x u ,y u ) is the predicted position of the person being located.

[0082] In order to verify the indoor positioning performance of this algorithm, it was tested in an experimental field and compared with other traditional positioning algorithms.

[0083] The experimental site is an office with a length of 14 meters and a width of 12 meters, with obstacles such as tables, chairs, and lockers. The number of access points is 3, and the reference point interval is 1.2 meters.

[0084] In the offline stage, 30 location fingerprints are collected at a frequency of 5 Hz for each reference point to build a location fingerprint database; in the online stage, the experimenter moves freely with the collection device in hand, collects a location fingerprint every 1 second, and records the location coordinates of the experimenter to build a test set.

[0085] The comparison algorithms in this experiment are: KNN, WKNN, RF, and SVM.

[0086] Experiments show that the average positioning error of this algorithm is 0.81m, which is better than other traditional positioning algorithms. The cumulative distribution function of the positioning error of different positioning algorithms is shown in Figure 2 , box plot see Figure 3 .

[0087] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A WiFi fingerprint positioning method based on the minimum coverage circle of reference points, characterized in that: This method uses the minimum coverage circle of reference points to determine whether there are reference points that are far away but have similar location fingerprints. When this abnormal phenomenon occurs, the selected location fingerprint is filtered with the help of the location of the person being located at the previous moment to reduce the impact of reference points that are far away but have similar location fingerprints on positioning accuracy. The method comprises at least the following steps: S1, calculate the fusion distance between the location fingerprint uploaded by the locatee and all the location fingerprints in the location fingerprint database; S2, arrange the fusion distances in ascending order and calculate the difference between adjacent fusion distances; S3, find the index value K of the maximum difference; S4, rearrange the location fingerprints in the location fingerprint database according to the principle of fusion distance from small to large, and select the first K location fingerprints; S5, query the reference points corresponding to the K position fingerprints, calculate the area S of the minimum coverage circle, and set the threshold λ; S6, if S<λ, weighted positioning is performed based on the K position fingerprints; if S≥λ, the position fingerprints are filtered in combination with the position of the person being positioned at the previous moment, and then weighted positioning is performed.

2. The WiFi fingerprint positioning method based on the minimum coverage circle of reference points according to claim 1 is characterized in that: In step S1, the fusion distance between the location fingerprint uploaded by the locatee and all the location fingerprints in the location fingerprint database is calculated as follows: FD(FP i ,FP j )=ωED(FP i ,FP j )+(1-ω)CD(FP i ,FP j ) Among them, FP i and FP j are two different location fingerprints, ED(FP i ,FP j ) is their Euclidean distance, CD(FP i ,FP j ) is their cosine distance, ω is the weight of the Euclidean distance, 0<ω<1, FD(FP i ,FP j ) is their fusion distance.

3. The WiFi fingerprint positioning method based on the minimum coverage circle of reference points according to claim 1 is characterized in that: In step S2, all fusion distances are arranged in ascending order as follows: (FD1,FD2,...,FD n ) Where n is the number of location fingerprints in the location fingerprint database; And the difference between two adjacent fusion distances is calculated as follows: dif i,j =FD j -FD i (dif 1,2 ,dif 2,3 ,...,dif n-1,n )。 4. The WiFi fingerprint positioning method based on the minimum coverage circle of reference points according to claim 1 is characterized in that: In step S3, the index value of the maximum difference is found as follows: K=argmax(dif 1,2 ,dif 2,3 ,...,dif n-1,n ) And, the index value starts from 1.

5. The WiFi fingerprint positioning method based on the minimum coverage circle of reference points according to claim 1 is characterized in that: In step S5, the threshold λ is calculated as follows: λ=πl 2 Where l is the minimum spacing between reference points.

6. The WiFi fingerprint positioning method based on the minimum coverage circle of reference points according to claim 1 is characterized in that: In step S6, if S<λ, weighted positioning is performed according to the following formula based on the K location fingerprints selected from the location fingerprint database: Among them, FP u is the location fingerprint uploaded by the located person, FPi is the location fingerprint in the i-th selected location fingerprint database, (x i ,y i ) is the location where it is collected, (x u ,y u ) is the predicted position of the located person; If S ≥ λ, the selected location fingerprint is filtered in combination with the location of the person being located at the last moment to delete the location fingerprint whose collection location is far from the location of the person being located; the filtering condition is: ED(P pre ,P i )>μ Among them, P pre is the position of the person being located predicted in the last positioning, P i is the collection location of the i-th location fingerprint, and μ is the maximum walking distance of the person being located within the positioning interval; the location fingerprints that meet this condition are deleted, and weighted positioning is performed according to the following formula based on the filtered K′ location fingerprints: Among them, FP u is the location fingerprint uploaded by the located person, FPi is the location fingerprint in the i-th selected location fingerprint database, (x i ,y i ) is the location where it is collected, (x u ,y u ) is the predicted position of the person being located.

7. A WiFi fingerprint positioning system based on a minimum coverage circle of reference points, characterized in that: At least: Access point, used to send WiFi signals; Location fingerprint collection device, a mobile communication device used to collect the RSS of the access point WiFi signal; The positioning server stores the location fingerprint database constructed in the offline stage, has a built-in program for executing the WiFi fingerprint positioning method based on the minimum coverage circle of reference points as described in any one of claims 1 to 6, receives the location fingerprint sent by the person being positioned and predicts his position.

Images