Indoor positioning method based on RSSI and Gaussian LOF outlier compensation

Abstract

The invention relates to an indoor positioning method based on RSSI and Gaussian LOF outlier compensation. The RSSI of a point to be positioned is obtained by deploying beacons; modifying the positions of beacons to simulate fault conditions such as beacon movement possibly occurring in actual production and life; obtaining positioning information of a to-be-positioned point by using a plurality of beacon coordinates and the RSSI, performing normalization processing, and determining an index and coordinates of an outlier by using a Gaussian LOF algorithm; after outliers are removed, positioning information is recovered through reverse normalization, and the maximum adjacent coordinate pair is searched; new coordinates with the same number as the outliers are uniformly inserted into the maximum adjacent coordinate pair, positioning information compensation after the outliers are removed is completed, and a final positioning result is obtained and output. According to the method provided by the invention, the problem of inaccurate positioning caused by beacon position movement or beacon signal blocking is solved, and higher-precision and more reliable indoor positioning is realized.

Description

Technical Field

[0001] The present invention relates to the field of positioning technology, and in particular to an indoor positioning method based on RSSI and Gaussian LOF outlier compensation. Background Art

[0002] As one of the core supporting technologies and important applications of the Internet of Things, wireless positioning technology and location-based services have always been widely concerned by the industry. Although satellite positioning technology represented by the Global Positioning System can provide high-precision positioning services in open outdoor environments, in closed environments such as indoors and underground, satellite positioning signals are blocked and cannot provide effective positioning services. Therefore, indoor positioning technology has emerged and is widely used in equipment positioning, industrial safety, and site inspections, effectively making up for the shortcomings of satellite positioning technology.

[0003] Most indoor positioning technologies require the deployment of beacon nodes, and the deployment topology of beacon nodes has a great impact on the determination of the location information of the located point during the actual positioning process. In actual production and life, the signal of the beacon node may be blocked and reflected, and even the location of the beacon may be illegally moved, thus affecting the signal reception strength RSSI.

[0004] The local outlier factor (LOF) algorithm introduces local density. Without making too many assumptions about the data distribution, it can effectively quantify the degree of anomaly of each data point and accurately search for outliers whose local density is significantly different from that of adjacent points. Summary of the invention

[0005] In view of this, in order to achieve higher precision, more reliable and self-correcting indoor positioning, the present invention provides an indoor positioning method based on RSSI and Gaussian LOF outlier compensation, comprising the following steps:

[0006] S1. Deploy beacons to obtain the signal receiving strength RSSI sequence;

[0007] S2. Obtain positioning information using RSSI and the coordinates of the moved beacon position;

[0008] S201. The distance d between the point to be located and the beacon is obtained by reverse calculation using the RSSI formula:

[0009] F RSSI (d) = F RSSI (d0)-10ηlog 10 (d)+Xs

[0010] Where d represents the distance between the point to be located and the beacon, F RSSI (d0) represents the RSSI value of the known reference distance, η represents the path loss factor, and Xs represents the shielding attenuation;

[0011] S202, sorting the RSSI sequence obtained in step S1 in descending order, selecting the first M RSSI values ​​and corresponding beacons as the basis for positioning, and calculating the distance d from the M corresponding beacons to the point to be positioned;

[0012] S203, using the least square method to calculate the coordinates (x, y) of the point to be located;

[0013] S3, normalize the positioning information, and use the Gaussian LOF algorithm to determine the index and coordinates of the outlier points;

[0014] S301, respectively calculate the distance from each positioning point to other positioning points, and express it with a symmetric distance matrix D:

[0015]

[0016] Where n represents the number of positioning points, and each element value d of the symmetric distance matrix D ij Represents the distance between the two positioning points corresponding to the i-th row and the j-th column;

[0017] S302, obtaining the Kth neighboring distance d of each positioning point through the symmetric distance matrix D K (P), K neighbors are N K (P);

[0018] S303, calculate the reachable distance P K (O i ,P),i∈(1,2,…,K), where point O i For point P in K neighborhood N K Points within (P);

[0019] S304. Calculate the local reachable density L using Gaussian kernel RD , the formula is as follows:

[0020] G(O i ,P)=exp(-γR K (O i ,P))

[0021]

[0022] Among them, G() represents the Gaussian kernel function, and γ represents a positive parameter used to control the width of the kernel function;

[0023] S305, using local reachability density L RD The local outlier factor is calculated by averaging the ratio of the local reachability density of all points in the K domain of the i-th anchor point to the local reachability density of the i-th anchor point. The formula is as follows:

[0024]

[0025] Among them, S i represents the local outlier factor value of the i-th positioning point, and N(i) represents the set of all positioning points in the K neighborhood of the i-th positioning point;

[0026] S306. Using the mean and standard deviation of the local outlier factor, a threshold θ is set. The formula is as follows:

[0027] θ=μ+m*σ

[0028] Among them, μ represents the mean of the local outlier factor, σ represents the standard deviation of the local outlier factor, and m represents the fusion ratio of the mean and standard deviation set according to the on-site environment;

[0029] Points exceeding the threshold are determined as outliers and the corresponding index and positioning coordinates are output;

[0030] S4, after removing outliers according to the index, perform denormalization, restore the positioning information, and search for the adjacent maximum distance coordinate pair;

[0031] S5. Evenly insert new coordinates into the adjacent maximum distance coordinate pairs, compensate the positioning information after removing the outliers, and obtain the final positioning result;

[0032] Linear interpolation is performed based on the adjacent maximum distance coordinate pair and the number of outlier indexes output in step S306, the coordinate data at both ends of the adjacent maximum distance coordinate pair are extracted, and positioning compensation points equal to the number of outliers are uniformly inserted from the horizontal coordinate and the vertical coordinate to obtain the final positioning result and output the positioning information.

[0033] Preferably, step S1 further comprises the following steps:

[0034] S101, deploying beacons is to evenly deploy a number of beacons along both sides of an indoor L-shaped corridor;

[0035] S102, collecting all RSSI values ​​of the points to be located, and sorting them from large to small;

[0036] S103, exchanging or moving some beacons originally arranged on both sides of the L-shaped corridor.

[0037] Preferably, the coordinates (x, y) of the point to be located are calculated using the least squares method:

[0038]

[0039] Among them, (x K ,y K) represents the beacon coordinates corresponding to the Mth RSSI value, d K Indicates the distance from the beacon corresponding to the Mth RSSI value to the point to be located.

[0040] Preferably, the reachable distance R K (O i ,P) represents point O i The K neighbor distance d K (P) and point O i The larger value of the distance to point P:

[0041] R K (O i ,P)=max{d K (O i ),d(O i ,P)}

[0042] Among them, point O i ,i∈(1,2,…,K) is a point in the K neighborhood of point P.

[0043] Preferably, the expression of the mean μ of the local outlier factor is as follows:

[0044]

[0045] The standard deviation expression of the local outlier factor is as follows:

[0046]

[0047] Among them, n represents the total number of points to be located, S i Represents the local outlier factor value of the i-th positioning point.

[0048] Preferably, outliers are removed from the positioning information according to the index of the outliers output in step S306, the positioning data is restored by anti-normalization, the distance from each positioning point to other positioning points is calculated, and represented by a symmetric distance matrix D′:

[0049]

[0050] Among them, n′ represents the number of positioning points after removing outliers, and each element value d′ of the symmetric distance matrix D′ ij Represents the distance and index between the two positioning points corresponding to the i-th row and the j-th column;

[0051] Sort all the nearest distances to get the maximum nearest distance and index and output the coordinate pair with the maximum nearest distance.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The present invention improves the local outlier factor LOF through the Gaussian kernel. Without making too many assumptions on the data distribution, it can effectively quantify the degree of anomaly of each data point, accurately search for outliers whose local density is obviously different from the local density of adjacent points, and compensate for the positioning data after eliminating the outliers, thereby achieving higher-precision and more reliable indoor positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a flow chart of the present invention;

[0055] Figure 2 It is a simulation diagram in an embodiment of the present invention, wherein the blue hollow dots represent normal positioning points, the red solid dots represent outliers detected by the present invention, and the green solid dots represent positioning points compensated by the present invention. DETAILED DESCRIPTION

[0056] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0057] The present invention provides an indoor positioning method based on RSSI and Gaussian LOF outlier compensation. In this embodiment, it is carried out for an L-shaped corridor with a length of 50 meters and a width of 5 meters in an indoor plane of 100 meters * 100 meters;

[0058] The following steps are involved:

[0059] S1. Deploy beacons to obtain the signal receiving strength RSSI sequence;

[0060] S101, deploying beacons is to evenly deploy 100 beacons on both sides of the indoor L-shaped corridor;

[0061] S102, each beacon takes itself as the origin, transmits its own number information and location information to the surroundings, collects all RSSI values ​​of the points to be located, and sorts them from large to small;

[0062] S103, artificially randomly exchange or move some beacons originally arranged on both sides of the L-shaped corridor, to simulate situations that may occur in actual production activities;

[0063] S2. Obtain positioning information using RSSI and the coordinates of the moved beacon position;

[0064] S201. The distance d between the point to be located and the beacon is obtained by reverse calculation using the RSSI formula:

[0065] F RSSI (d) = F RSSI (d0)-10ηlog10 (d)+Xs

[0066] Where d represents the distance between the point to be located and the beacon, F RSSI (d0) represents the RSSI value of the known reference distance. In this embodiment, F RSSI (d0) = -40 dBm, η represents the path loss factor. In this embodiment, the path loss factor η of each beacon is set to a value in a random distribution with a mean of 2 and a standard deviation of 0.2. Xs represents occlusion attenuation, which represents the change in signal strength caused by obstacles in the environment. In this embodiment, the occlusion attenuation Xs is set to a Gaussian distribution with a mean of 0 and a standard deviation of 0.5;

[0067] S202, sort the RSSI sequence obtained in step S1 in descending order, select the first M RSSI values ​​and corresponding beacons as the basis for positioning. In this embodiment, M is 4, and the distance d from the 4 corresponding beacons to the point to be positioned is calculated;

[0068] S203, using the least square method to calculate the coordinates (x, y) of the point to be located, the expression is as follows:

[0069]

[0070] Among them, (x i ,y i ),i∈(1,2,3,4) represents the beacon coordinates corresponding to the ith RSSI value, d i Indicates the distance from the beacon corresponding to the i-th RSSI value to the point to be located;

[0071] Write the coordinates of the points to be located in matrix form and make a simple transformation to get:

[0072] X=(A T A) -1 A T b

[0073] in:

[0074] X=(x,y)

[0075]

[0076] S3, normalize the positioning information, and use the Gaussian LOF algorithm to determine the index and coordinates of the outlier points;

[0077] In this embodiment, a maximum-minimum normalization method is used;

[0078] S301, respectively calculate the distance from each positioning point to other positioning points, and express it with a symmetric distance matrix D:

[0079]

[0080] Where n represents the number of positioning points, and each element value d of the symmetric distance matrix D ij Represents the distance between the two positioning points corresponding to the i-th row and the j-th column;

[0081] S302, obtaining the Kth neighboring distance d of each positioning point through the symmetric distance matrix D K (P), K neighbors are N K (P), in this embodiment, K is set to 5, and the formula is as follows:

[0082] N5(P)={d(P,O i )≤d5(P)}

[0083] Point P is the center of the neighborhood circle, O i For all points in the neighborhood, d(P,O i ) represents point P to point O i In this embodiment, all distances are Euclidean distances;

[0084] S303, calculate the reachable distance R5 (0 i ,P),i∈(1,2,…,5), reachable distance R5(O i ,P) represents point O i The K=5 neighboring distance d5(P) and point O i The larger value of the distance to point P:

[0085] R5(O i ,P)=max{d5(O i ),d(O i ,P)}

[0086] Among them, point O i , i∈(1,2,…,5) is a point in the K=5 neighborhood of point P, d5(O i ) represents point O i The proximity distance, d(O i ,P) represents point O i The distance to point P;

[0087] S304. Calculate the local reachable density L using Gaussian kernel RD , the formula is as follows:

[0088] G(O i ,P)=exp(-γR5(O i ,P))

[0089]

[0090] Wherein, K() represents a Gaussian kernel function, γ represents a positive parameter used to control the width of the kernel function, and in this embodiment, γ is set to 0.25;

[0091] S305, using local reachability density L RD Calculate the local outlier factor, the formula is as follows:

[0092]

[0093] Among them, S i represents the local outlier factor value of the ith positioning point, N(i) represents all the positioning points in the K=5 neighborhood of the ith positioning point, and L RD (j) represents the local reachable density obtained in step S304, and whether the i-th positioning point is an outlier is determined by averaging the ratio of the local reachable density of all points in the K=5 neighborhood of the i-th positioning point to the local reachable density of the i-th positioning point;

[0094] S306, using the local outlier factor S i The mean and standard deviation of the threshold θ are set as follows:

[0095] θ=μ+m*σ

[0096] Wherein, μ represents the mean of the local outlier factor, σ represents the standard deviation of the local outlier factor, and m represents the fusion ratio of the mean and the standard deviation set according to the on-site environment. In this embodiment, m is set to 2;

[0097] The expression of the mean μ of the local outlier factor is as follows:

[0098]

[0099] The standard deviation expression of the local outlier factor is as follows:

[0100]

[0101] Where n represents the total number of points to be located. In this embodiment, n is set to 160. i Represents the local outlier factor value of the i-th positioning point;

[0102] Points exceeding the threshold are determined as outliers and the corresponding index and positioning coordinates are output;

[0103] S4, after removing outliers according to the index, perform denormalization, restore the positioning information, and search for the adjacent maximum distance coordinate pair;

[0104] The outliers are removed from the positioning information according to the index of the outliers output in step S306, the positioning data is restored by anti-normalization, the distance from each positioning point to other positioning points is calculated, and represented by a symmetric distance matrix D′:

[0105]

[0106] Among them, n′ represents the number of positioning points after removing outliers, such as Figure 2 As shown, in this embodiment, the number of outliers removed is 3, so n′=157;

[0107] Each element value d′ of the symmetric distance matrix D′ ij Represents the distance and index between the two positioning points corresponding to the i-th row and the j-th column;

[0108] Sort all the nearest distances to get the maximum nearest distance and index and output the coordinate pair with the maximum nearest distance;

[0109] S5. Evenly insert new coordinates into the adjacent maximum distance coordinate pairs, compensate the positioning information after removing the outliers, and obtain the final positioning result;

[0110] Linear interpolation is performed based on the adjacent maximum distance coordinate pair and the number of outlier indexes output in step S306, the coordinate data at both ends of the adjacent maximum distance coordinate pair are extracted, and positioning compensation points equal to the number of outliers are uniformly inserted from the horizontal coordinate and the vertical coordinate to obtain the final positioning result and output the positioning information.

[0111] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. An indoor positioning method based on RSSI and Gaussian LOF outlier compensation, characterized in that: The following steps are involved: S1. Deploy beacons to obtain the signal receiving strength RSSI sequence; S2, using RSSI and the coordinates of the moved beacon to obtain positioning information; S201. The distance d between the point to be located and the beacon is obtained by reverse calculation using the RSSI formula: F RSSI (d)=F RSSI (d0)-10ηlog 10 (d)+Xs Where d represents the distance between the point to be located and the beacon, F RSSI (d0) represents the RSSI value of the known reference distance, η represents the path loss factor, and Xs represents the shielding attenuation; S202, sorting the RSSI sequence obtained in step S1 in descending order, selecting the first M RSSI values ​​and corresponding beacons as the basis for positioning, and calculating the distance d from the M corresponding beacons to the point to be positioned; S203, using the least square method to calculate the coordinates (x, y) of the point to be located; S3, normalize the positioning information, and use the Gaussian LOF algorithm to determine the index and coordinates of the outlier points; S301, respectively calculate the distance from each positioning point to other positioning points, and express it with a symmetric distance matrix D: Among them, n represents the number of positioning points, and each element value d of the symmetric distance matrix D ij Represents the distance between the two positioning points corresponding to the i-th row and the j-th column; S302, obtaining the Kth neighboring distance d of each positioning point through the symmetric distance matrix D K (P), K neighbors are N K (P); S303, calculate the reachable distance R K (O i ,P),i∈(1,2,…,K), where point O i For point P in K neighborhood N K Points within (P); S304. Calculate the local reachable density L using Gaussian kernel RD , the formula is as follows: G(O i ,P)=exp(-γR K (O i ,P)) Among them, G() represents the Gaussian kernel function, and γ represents a positive parameter used to control the width of the kernel function; S305, using local reachability density L RD The local outlier factor is calculated by averaging the ratio of the local reachability density of all points in the K domain of the i-th anchor point to the local reachability density of the i-th anchor point. The formula is as follows: Among them, S i represents the local outlier factor value of the i-th positioning point, and N(i) represents the set of all positioning points in the K neighborhood of the i-th positioning point; S306, using the local outlier factor S i The mean and standard deviation of the threshold θ are set as follows: θ=μ+m*σ Among them, μ represents the mean of the local outlier factor, σ represents the standard deviation of the local outlier factor, and m represents the fusion ratio of the mean and standard deviation set according to the on-site environment; Points exceeding the threshold are determined as outliers and the corresponding index and positioning coordinates are output; S4, after removing outliers according to the index, perform denormalization, restore the positioning information, and search for the adjacent maximum distance coordinate pair; S5. Evenly insert new coordinates into the adjacent maximum distance coordinate pairs, compensate the positioning information after removing the outliers, and obtain the final positioning result; Linear interpolation is performed based on the adjacent maximum distance coordinate pair and the number of outlier indexes output in step S306, the coordinate data at both ends of the adjacent maximum distance coordinate pair are extracted, and positioning compensation points equal to the number of outliers are uniformly inserted from the horizontal coordinate and the vertical coordinate to obtain the final positioning result and output the positioning information.

2. The indoor positioning method based on RSSI and Gaussian LOF outlier compensation according to claim 1, characterized in that: Step S1 also includes the following steps: S101, deploying beacons means evenly deploying a number of beacons along both sides of an indoor L-shaped corridor; S102, collecting all RSSI values ​​of the points to be located and sorting them from large to small; S103, exchanging or moving some beacons originally arranged on both sides of the L-shaped corridor.

3. The indoor positioning method based on RSSI and Gaussian LOF outlier compensation according to claim 1, characterized in that: The coordinates (x, y) of the point to be located are calculated using the least squares method: Among them, (x M ,y M ) represents the beacon coordinates corresponding to the Mth RSSI value, d M Indicates the distance from the beacon corresponding to the Mth RSSI value to the point to be located.

4. The indoor positioning method based on RSSI and Gaussian LOF outlier compensation according to claim 1, characterized in that: The reachable distance R K (O i ,P) represents point O i The K neighbor distance d K (P) and point O i The larger value of the distance to point P: P K (O i ,P)=max{d K (O i ),d(O i ,P)} Among them, point O i ,i∈(1,2,…,K) is a point in the K neighborhood of point P.

5. The indoor positioning method based on RSSI and Gaussian LOF outlier compensation according to claim 1, characterized in that: The expression of the mean μ of the local outlier factor is as follows: The standard deviation expression of the local outlier factor is as follows: Among them, n represents the total number of points to be located, S i Represents the local outlier factor value of the i-th positioning point.

6. The indoor positioning method based on RSSI and Gaussian LOF outlier compensation according to claim 1, characterized in that: The outliers are removed from the positioning information according to the index of the outliers output in step S306, the positioning data is restored by anti-normalization, the distance from each positioning point to other positioning points is calculated, and represented by a symmetric distance matrix D′: Among them, n′ represents the number of positioning points after removing outliers, and each element value d′ of the symmetric distance matrix D′ ij Represents the distance and index between the two positioning points corresponding to the i-th row and the j-th column; Sort all the nearest distances to get the maximum nearest distance and index and output the coordinate pair with the maximum nearest distance.

Images