A three-dimensional position fingerprint database construction method and an electronic device

Abstract

The application discloses a three-dimensional position fingerprint database construction method and electronic equipment. The method comprises the following steps: establishing a three-dimensional coordinate system, taking the three-dimensional coordinate axes as the reference, and performing three-dimensional grid division on a target region; taking the grid corner points as sampling points, collecting RSSI observation values of the sampling points, recording position coordinates of the sampling points, and constructing a three-dimensional position fingerprint database of the sampling points. Interpolation points are selected, for each interpolation point, a region constraint is introduced respectively, and a weight data set of the sampling points to the interpolation point meeting the region constraint is established. The weight data set is updated based on the AP point position constraint; RSSI prediction values of the interpolation points are calculated based on the updated weight data set, and a three-dimensional position fingerprint database of the interpolation points is constructed. The three-dimensional position fingerprint database of the sampling points and the three-dimensional position fingerprint database of the interpolation points are combined to obtain a final three-dimensional position fingerprint database. The application has high interpolation precision and low data acquisition cost.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of indoor positioning, in particular to a three-dimensional position fingerprint database construction method and electronic equipment. BACKGROUND

[0002] Nowadays, the location-based service (LBS) in the ubiquitous computing is a very promising business, which can widely support various applications requiring dynamic position information, and provide relatively accurate position information for information query, traffic management, medical aid, road assistance and navigation, etc., thus providing LBS for users has a huge market scale and good business prospects.

[0003] According to different applications of positioning technology, the indoor positioning system in China can be divided into passive personnel positioning based on online patrol, personnel positioning based on Wi-Fi, personnel positioning based on ultra-wideband (UWB) technology, personnel positioning based on radio frequency identification (RFID) technology and personnel positioning based on Bluetooth technology, etc. Among them, the position fingerprint positioning method based on Wi-Fi is most widely used due to its low cost, wide coverage, strong signal interference and other properties, and the terminal equipment carried by individuals can receive the received signal strength indication (RSSI) values from different access points (APs) at the same time, and it is low-cost to arrange Wi-Fi devices in places with large human flow, which is the best choice for indoor positioning.

[0004] When the personnel are in an indoor environment where mobile phone signals and GPS satellite positioning signals cannot be received, APs can be arranged at different heights in each area for positioning. However, due to different geographical environments, when using the traditional position fingerprint positioning algorithm, different sizes of offline position fingerprint data need to be collected to construct a large offline position fingerprint library, which requires a large amount of manual cost. SUMMARY

[0005] The present application aims to provide a three-dimensional position fingerprint database construction method and electronic equipment, which has high interpolation accuracy and low data acquisition cost.

[0006] The technical scheme adopted by the present application is:

[0007] On the one hand, the present application provides a three-dimensional position fingerprint database construction method, comprising:

[0008] Step 1, a three-dimensional coordinate system is established for the target area, and the target area is divided into three-dimensional grids (unit grid division);

[0009] The RSSI observation value at the sampling point is collected, and the position coordinates of the sampling point are recorded to construct a three-dimensional position fingerprint database of the sampling points.

[0010] Step 2, selecting interpolation points in the target area, introducing a region constraint for each interpolation point, and establishing a weight data set of the sampling points to the interpolation point based on the inverse distance weighted interpolation algorithm to meet the region constraint;

[0011] Step 3, updating the weight data set based on the AP point position constraint, calculating the RSSI prediction value of the interpolation point based on the updated weight data set, and constructing a three-dimensional position fingerprint database of the interpolation points;

[0012] Step 4, combining the three-dimensional position fingerprint database of the sampling points with the three-dimensional position fingerprint database of the interpolation points to obtain a final three-dimensional position fingerprint database (offline three-dimensional position fingerprint database).

[0013] As an implementation manner of the first aspect, in the step 1, the RSSI observation value is filtered first, and then the three-dimensional position fingerprint database of the sampling points is constructed based on the filtered data.

[0014] As an implementation manner of the first aspect, the step 1 includes:

[0015] S11, a three-dimensional coordinate system is randomly set in the target area, a three-dimensional grid is divided in the target area with the three-dimensional coordinate axes as the reference and l as the grid side length, a first grid corner point set C1 is formed by the grid corner points obtained by the division, and the grid corner points in the first grid corner point set C1 are set as the sampling points.

[0016] S12, the RSSI observation value of each sampling point is collected and saved;

[0017] S13, a three-dimensional position fingerprint database R of the sampling points is constructed o :

[0018] R o = {R i |i = 1, 2, …, n}

[0019] R i = [RSSI i1 , RSSI i2 , …, RSSI iV , x i , y i , z i ]

[0020] Wherein, n is the total number of sampling points, RSSI ivdenotes the RSSI observation value collected at the i-th sampling point for the v-th time, v = 1, 2, …, V, V is the number of data collection times, (x i ,y i ,z i ) is the coordinate of the i-th sampling point in the three-dimensional coordinate system.

[0021] As an implementation manner of the first aspect, the step 2 comprises:

[0022] For any interpolation point P inter , a region constraint is introduced, and a weight data set of the sampling points to the interpolation point P inter satisfying the region constraint is established based on the inverse distance weighted interpolation algorithm, comprising:

[0023] S21, a corresponding sampling point distance data set H of the interpolation point P inter is established:

[0024] H = {h i |i = 1, 2, …, n}

[0025]

[0026] wherein h i is the distance from the i-th sampling point to the interpolation point P inter ; (x o ,y o ,z o ) is the coordinate of the interpolation point P inter in the three-dimensional coordinate system;

[0027] S22, a spherical region constraint is introduced: a spherical region A1 is determined with the interpolation point P inter as the spherical center and r as the radius; wherein the radius r satisfies the following formula:

[0028] The distance data set of the sampling points in the spherical region A1 to the interpolation point P inter is updated and saved:

[0029] H * = {g j |j = 1, 2, …, m}

[0030]

[0031] wherein g j is the distance from the j-th sampling point in the spherical region A1 to the interpolation point, (x′ j ,y′ j ,z′ j ) is the coordinate of the j-th sampling point in the spherical region A1, and m is the number of sampling points in the spherical region A1.

[0032] S23, set the minimum number of sampling points M, determine whether m is greater than or equal to M, if the result is yes, continue step S24, otherwise expand the spherical region A1 radius, that is, r = r + l, and return to S22;

[0033] S24, based on H * The weight data set W of the sampling points in the spherical region A1 to the interpolation point P inter is established:

[0034] W = {W j |j = 1, 2,..., m}

[0035]

[0036] Wherein, u is the exponential parameter.

[0037] As an implementation form of the first aspect, the step 3 comprises:

[0038] For any interpolation point P inter , let its coordinate position in the three-dimensional coordinate system be (x o , y o , z o ), the following method is used to calculate the vth RSSI prediction value at the interpolation point P inter , including:

[0039] S31, determine the coordinates (x ap , y ap , z ap ) of the AP point in the target region in the three-dimensional coordinate system, and calculate the distance r inter from the interpolation point P * to the AP point;

[0040] S32, take the AP point in the target region as the center of the sphere, and r * as the radius to make a spherical surface, and calculate the shortest distance d from each sampling point in the spherical region A1 to the spherical surface:

[0041] d = {d j |j = 1, 2,..., m}

[0042]

[0043] S33, determine the position constraint weight λ of the AP point based on the following formula:

[0044] λ = {λ j |j = 1, 2,..., m}

[0045]

[0046] wherein u is any positive real number;

[0047] S34, calculating the interpolation point P based on the following formula inter The vth RSSI prediction value is:

[0048]

[0049] wherein RSSI jv is the vth RSSI observation value of the jth sampling point in the spherical area A1, RSSI v * represents the interpolation point P inter The vth RSSI prediction value.

[0050] As an implementation form of the first aspect, the step 3 further comprises:

[0051] constructing an interpolation point three-dimensional position database R * :

[0052] R * = {R k * | k = 1, 2, …, p}

[0053] R k * = [RSSI k1 * , RSSI k2 * , …, RSSI kV * , x k * , y k * , z k * ]

[0054] wherein p represents the total number of interpolation points, R k * represents the RSSI prediction value and coordinate information of the kth interpolation point, RSSI kv * represents the vth RSSI prediction value of the kth interpolation point, RSSI kv * , (x k * , y k * , z k * ) is the coordinate of the kth interpolation point.

[0055] As an implementation form of the first aspect, the step 4 comprises combining the sampling point three-dimensional position fingerprint database Ro a three-dimensional location fingerprint database R * , to obtain a final three-dimensional location fingerprint database R = {R o ∪R *}.

[0056] In a second aspect, the present application provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor implements the method.

[0057] In a third aspect, the present application provides a computer readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the processor implements the method.

[0058] Advantages:

[0059] Compared with the traditional location fingerprint method, the present application improves the inverse distance weighted interpolation method (IDW) and fuses the AP position constraint, which has the following advantages:

[0060] 1. The improved inverse distance weighted interpolation method proposed in the present application has faster interpolation efficiency and higher interpolation accuracy than the traditional inverse distance weighted algorithm.

[0061] 2. The present application fuses the AP point position constraint and the improved inverse distance weighted interpolation method, which can reduce more than 50% of the artificial offline collection workload and reduce the pre-collection cost.

[0062] 3. Using the fingerprint library proposed in the present application for positioning test can improve the average positioning accuracy and reduce the overall error. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 is a flowchart of the construction of a three-dimensional location fingerprint database of an embodiment of the present application.

[0064] Figure 2 is a flowchart of the interpolation algorithm provided by an embodiment of the present application.

[0065] Figure 3 is a comparison chart of the RSSI values of the interpolation point three-dimensional location fingerprint database and the original fingerprint library constructed by an embodiment of the present application.

[0066] Figure 4 is a top view of the experimental scene of the embodiment of the present application, and the cross mark represents the AP position, with the unit of millimeter. DETAILED DESCRIPTION

[0067] The technical solutions of the present application will be described clearly and completely below in connection with the drawings and specific embodiments. Obviously, the described embodiments are only some of the embodiments of the present application, but not all.

[0068] Figure 1 The flow chart of constructing the three-dimensional location fingerprint database provided by an embodiment of the present application, Figure 2 The flow chart of the interpolation method provided by an embodiment of the present application. Referring to Figure 1 、 Figure 2 The three-dimensional location fingerprint database construction method provided by the embodiment of the present application can include a data collection stage and an interpolation stage, wherein the data collection stage can include the following step 1, and the interpolation stage can include the following steps 2 to 4, and the specific implementation process is as follows:

[0069] Step 1, establishing a three-dimensional coordinate system for the target area, and performing three-dimensional grid division (unit grid division) on the target area;

[0070] Taking the grid corner points in the target area as sampling points, collecting the received signal strength value (RSSI) observation value of the sampling points, and recording the position coordinates of the sampling points, a three-dimensional location fingerprint database of the sampling points is constructed;

[0071] Step 2, selecting interpolation points in the target area, introducing regional constraints, and establishing a weight data set from the sampling points to the interpolation points based on the inverse distance weighted interpolation algorithm to meet the regional constraints;

[0072] Step 3, updating the above weight data set based on the AP point position constraint; calculating the RSSI prediction value of the interpolation points based on the updated weight data set, and constructing a three-dimensional location fingerprint database of the interpolation points;

[0073] Step 4, combining the three-dimensional location fingerprint database of the sampling points with the three-dimensional location fingerprint database of the interpolation points to obtain the final three-dimensional location fingerprint database (offline three-dimensional location fingerprint database).

[0074] In an embodiment, in the step 1, the RSSI observation value can be filtered first, and then the three-dimensional location fingerprint database of the sampling points is constructed based on the filtered data.

[0075] Figure 3 is the RSSI comparison chart of the interpolation point three-dimensional location fingerprint database constructed by the present application and the original fingerprint library, Figure 4 is the experimental scene diagram of the embodiment of the present application, and the target area in the embodiment is a certain two-story 6m high indoor building, and the horizontal area is 3.3x4.8=15.84m 2 .

[0076] The following will be described in connection with Figure 4The experimental scene shown is used to illustrate the scheme of the present application.

[0077] Step 1 is implemented as follows:

[0078] S11, in the initialization stage, a three-dimensional coordinate system is randomly set in the target area, and the target area is divided into a three-dimensional grid with the three-dimensional coordinate axes as the reference and l as the grid side length (unit length). The grid corner points obtained by the division form a first grid corner point set C1, and the grid corner points in the first grid corner point set C1 are set as sampling points.

[0079] P i = (x i ,y i ,z i )

[0080] where P i is the coordinates of the i-th sampling point in the three-dimensional coordinate system.

[0081] In an embodiment, the grid side length is set to l = 1 m, and the grid size is 1 × 1 × 1 m 3 . For the target area in the experimental scene shown, the number of sampling points is 3 × 4 × 2 = 24. Figure 4

[0082] S12, the RSSI observation value at each sampling point is collected by a sensor module (which can be a smart phone carrying a corresponding sensor module) and saved in a database.

[0083] For example, the RSSI observation value of the i-th sampling point can be saved as:

[0084] R i = [RSSI i1 ,RSSI i2 ,…,RSSI iV ,x i ,y i ,z i ]

[0085] where RSSI iv represents the RSSI observation value collected at the i-th sampling point for the v-th time, v = 1, 2, …, V, and V is the number of data collection times.

[0086] In this experiment, V = 16.

[0087] S13, a three-dimensional location fingerprint database R o is constructed:

[0088] R o = {R i |i = 1, 2, …, n} ​

[0089] wherein n is the total number of sampling points, and in this example n = 24.

[0090] The step 2 comprises:

[0091] Selecting interpolation points in the target region to construct an interpolation point set.

[0092] For each interpolation point in the interpolation point set C3, introduce a region constraint, and based on the inverse distance weighted interpolation algorithm, establish a weight data set of the sampling points to the interpolation point that satisfies the region constraint. Figure 4 As shown in the experimental scene of the target region, the position of the interpolation point on the plane is as shown in the transparent triangular point, that is, the target region is divided into a three-dimensional grid with a grid side length of l / 2, the grid corner points obtained by the division form a second grid corner point set C2, and the interpolation point set C3 = the second grid corner point set C2 - the first grid corner point set C1 is determined.

[0093] For each interpolation point in the interpolation point set C3, introduce a region constraint, and based on the inverse distance weighted interpolation algorithm, establish a weight data set of the sampling points to the interpolation point that satisfies the region constraint.

[0094] In an embodiment, for any interpolation point P inter , let its coordinates in the three-dimensional coordinate system be (x o , y o , z o ), introduce a region constraint, and based on the inverse distance weighted interpolation algorithm, establish a weight data set of the sampling points to the interpolation point P inter that satisfies the region constraint, which can include:

[0095] S21, establish a sampling point distance data set corresponding to the interpolation point P inter by the following formula:

[0096] H = {h i | i = 1, 2, …, n}

[0097]

[0098] wherein h i is the distance from the i-th sampling point to the interpolation point P inter .

[0099] S22, introduce a spherical region constraint: take the interpolation point P inter as the center of the sphere, and r as the radius to determine a spherical region A1, that is, the spherical region A1 corresponds to the following spherical surface equation:

[0100] (x - x o ) 2 + (y - y o ) 2 + (z - z o ) 2 = r 2

[0101] wherein the radius r satisfies the following formula:

[0102] updating and saving the distance dataset of the sampling points in the spherical region A1 to the interpolation point P inter :

[0103] H * = {g j |j = 1, 2, …, m}

[0104]

[0105] wherein g j is the distance from the jth sampling point in the spherical region A1 to the interpolation point, (x' j , y' j , z' j ) is the coordinate of the jth sampling point in the spherical region A1, and m is the number of sampling points in the spherical region A1.

[0106] S23, setting the minimum number of sampling points M, judging whether m is greater than or equal to M, if the result of the judgment is yes, continuing step S24, otherwise expanding the radius of the spherical region, i.e. r = r + l, and returning to S22.

[0107] The minimum number of sampling points M can be set to 4-8, and in the present example, the minimum number of sampling points M is set to 4.

[0108] S24, based on H * , establishing the weight dataset W of the sampling points in the spherical region A1 to the interpolation point P inter :

[0109] W = {W j |j = 1, 2, …, m}

[0110]

[0111] wherein u is an exponential parameter, which takes any positive real number, and in the present example, the value of u can be set to 2.

[0112] In one embodiment, the step 3 can include:

[0113] For any interpolation point P inter , let its coordinate position in the three-dimensional coordinate system be (x o , y o , z o ), the following method can be used to calculate the vth RSSI prediction value at the interpolation point P inter , including:

[0114] S31, determine the coordinates (x ap ,y ap ,z ap ) of the AP point in the target area in the three-dimensional coordinate system, and calculate the distance r inter from the interpolation point P * to the AP point, as follows:

[0115]

[0116] S32, take the AP point in the target area as the center of the sphere and r * as the radius to make a spherical surface, and calculate the shortest distance d (the absolute value of the difference between the distance from the sampling point to the AP point and the distance from the interpolation point to the AP point) from each sampling point in the spherical region A1 to the spherical surface:

[0117] d = {d j |j = 1, 2,..., m}

[0118]

[0119] S33, calculate the AP point position constraint weight λ based on the following formula:

[0120] λ = {λ j |j = 1, 2,..., m}

[0121]

[0122] where u is any positive real number, and in this example, u is 2.

[0123] Based on the AP point position constraint weight λ, the weight data set can be updated to W':

[0124] W' = {W j × λ j |j = 1, 2,..., m};

[0125] S34, calculate the RSSI prediction value of the interpolation point P inter for the vth time based on the following formula:

[0126]

[0127] where RSSI jv is the RSSI observation value of the jth sampling point in the spherical region A1 for the vth time, and RSSI v * represents the RSSI prediction value of the interpolation point P inter for the vth time, where v = 1, 2,..., V.

[0128] Let the RSSI prediction value of the kth interpolation point for the vth time be RSSI kv* The RSSI prediction value and coordinate information of the kth interpolation point can be recorded as:

[0129] R k * = [RSSI k1 * ,RSSI k2 * ,…,RSSI kv * ,x k * ,y k * ,z k * ]

[0130] where (x k * ,y k * ,z k * ) is the coordinate of the kth interpolation point.

[0131] By traversing each interpolation point, the RSSI prediction value of all interpolation points is calculated through the above steps. Combined with the coordinate information of the interpolation points, the interpolation point three-dimensional location database R * can be constructed.

[0132] R * = {R k * | k = 1, 2, …, p}

[0133] where p represents the total number of interpolation points.

[0134] Step 4 can include:

[0135] Combined with the sampling point three-dimensional location fingerprint database R o and the interpolation point three-dimensional location fingerprint database R * , the final (offline) three-dimensional location fingerprint database R = {R o ∪ R *} is established.

[0136] The traditional inverse distance weight method relies on the power value of the inverse distance, and only controls the influence of the known points on the interpolation based on the distance from the reference sampling point to the output point, although the calculation speed is improved, but the inverse distance weight formula is not related to any actual physical process, so it cannot be determined whether the specific power value is too large, which may generate an error result. The technical solution provided by the above embodiment of the present application improves the traditional inverse distance weight method, adds spatial distribution constraints and AP point position constraints of the sampling points, improves the calculation efficiency of the traditional inverse distance weight interpolation method, and the calculated interpolation is closer to the actual collected data than the traditional interpolation method; also can reduce the large amount of manual cost required by the traditional location fingerprint method when positioning in the indoor GPS-free environment, and improve the positioning accuracy.

[0137] Experimental verification:

[0138] The RSSI prediction error obtained by the fingerprint database construction method based on the improved inverse distance weight interpolation method (i.e. the inverse distance weight interpolation method with regional constraints) + AP point position constraint weight data set acquisition method, the fingerprint database construction method based on the traditional inverse distance weight method weight data set acquisition method, the fingerprint database construction method based on the improved inverse distance weight interpolation method weight data set acquisition method, and the fingerprint database construction method based on the traditional inverse distance weight interpolation method + AP point position constraint weight data set acquisition method, are shown in Table 1, unit (dBm):

[0139] Table 1: Comparison of interpolation error

[0140]

[0141] Figure 3 The error table obtained by the four fingerprint database calculation methods shows that the fingerprint database construction method proposed in the embodiment of the present application has the smallest error with the actual measured RSSI value, and the average error is 1.49 dBm, which is reduced by 0.4 dBm compared with the fingerprint database calculation error of 1.89 dBm based on the traditional inverse distance weight interpolation method, and the performance is improved by 21.16%; compared with the average errors of 1.59 dBm and 1.58 dBm of the fingerprint database based on the improved inverse distance weight interpolation method and the fingerprint database based on the traditional inverse distance weight interpolation method + AP point position constraint weight data set acquisition method, the average error of 1.49 dBm of the fingerprint database construction method proposed in the embodiment of the present application also has an advantage. Therefore, the interpolation method proposed in the embodiment of the present application has the smallest RSSI error and is more accurate than other methods.

[0142] At the same time, the error accumulation distribution probability comparison of the four interpolation methods is shown in Table 2:

[0143] Table II: Error accumulation probability of interpolation algorithm

[0144]

[0145] It can be seen from the error accumulation distribution obtained by the four fingerprint database calculation methods that, compared with the fingerprint database construction method based on the traditional inverse distance weight interpolation method, the fingerprint database construction method based on the improved inverse distance weight interpolation method, and the fingerprint database construction method based on the traditional inverse distance weight interpolation method + AP point position constraint, the RSSI error distribution calculated by the fingerprint database construction method of the application has the maximum probability in the three error ranges, which are 52.4%, 68.3%, and 82.5%, respectively. It can be known that the algorithm interpolation obtained by the application is similar to the original fingerprint database, and has great advantages.

[0146] In summary, the improved inverse distance weight interpolation algorithm based on the fusion of AP point position constraint of the application is more accurate than the traditional fingerprint database construction interpolation method, reduces the interpolation error, reduces the large space data calculation amount, reduces the artificial offline collection workload by more than 50%, reduces the pre-collection cost, and improves the average positioning accuracy when the fingerprint database of the application is used for interpolation prediction in the example.

Claims

1. A method of constructing a three-dimensional location fingerprint database, characterized by, The method comprises the following steps: Step 1, establishing a three-dimensional coordinate system, taking the three-dimensional coordinate axes as the reference, and performing three-dimensional grid division on the target area; Taking the grid corner points as the sampling points, collecting the received signal strength indication (RSSI) observation values at the sampling points, recording the position coordinates of the sampling points, and constructing a three-dimensional position fingerprint database of the sampling points; Step 2, selecting an interpolation point, introducing a region constraint for each interpolation point, and establishing a weight data set of the sampling points to the interpolation point based on the inverse distance weight interpolation algorithm to meet the region constraint; Step 3, updating the weight data set based on the AP point position constraint, calculating the RSSI prediction value of the interpolation point based on the updated weight data set, and constructing a three-dimensional position fingerprint database of the interpolation points; specifically comprising: For any interpolation point , let its coordinate position in a three-dimensional coordinate system be , the following method is used to calculate the th RSSI prediction value at the interpolation point , comprising: S31, determining the coordinates of the AP point in the three-dimensional coordinate system in the target region , and calculating the distance from the interpolation point to the AP point ;​ S32, taking the AP point in the target region as the center of a sphere, make a spherical surface with the radius, and calculate the shortest distance from each sampling point in the spherical region to the spherical surface : ； ； S33, determining the AP point position constraint weight based on the following formula : ； ； wherein is any positive real number; S34, calculate the interpolation point based on the following formula The first The second RSSI prediction value is: ； wherein, is the RSSI observation value of the i-th sampling point in the j-th spherical region, is the RSSI prediction value of the interpolation point in the j-th spherical region, the i-th sampling point in the j-th spherical region,​​​ Constructing an interpolation point three-dimensional position database : ； ； wherein, denotes the total number of interpolation points, denotes the RSSI prediction value and coordinate information of the th interpolation point, denotes the RSSI prediction value of the th interpolation point at the th time, denotes the coordinate of theth interpolation point;​ Step 4, combining the three-dimensional position fingerprint database of the sampling points with the three-dimensional position fingerprint database of the interpolation points to obtain the final three-dimensional position fingerprint data.

2. The three-dimensional position fingerprint database construction method according to claim 1, characterized by, In step 1, the RSSI observation values are first filtered, and then the three-dimensional position fingerprint database of the sampling points is constructed based on the filtered data.

3. The three-dimensional position fingerprint database construction method according to claim 1 or 2, characterized by, The step 1 comprises: S11, a three-dimensional coordinate system is randomly set in the target area, and a three-dimensional coordinate axis is taken as a reference, and a three-dimensional grid is divided for the target area with the three-dimensional coordinate axis as a reference, and a first grid corner point set C1 is formed by grid corner points obtained by the division, and the grid corner points in the first grid corner point set C1 are set as sampling points; S12, collecting and saving the RSSI observation values of each sampling point; S13, constructing a three-dimensional position fingerprint database of sampling points : ； ； wherein, is the total number of sampling points, denotes the RSSI observation value collected at the th sampling point at the th data collection, , is the number of data collections, is the coordinate of the th sampling point in the three-dimensional coordinate system.

4. The three-dimensional position fingerprint database construction method according to claim 3, characterized by, The step 2 comprises: For any interpolation point , introduce area constraints, based on inverse distance weighted interpolation algorithm to meet the area constraints of the sampling point to the interpolation point weight data set, including: S21, establishing the interpolation point corresponding sampling point distance data set : ； ； wherein is the distance from the th sampling point to the interpolation point ; and is the coordinate of the interpolation point in the three-dimensional coordinate system. S22, introduce a spherical region constraint: with interpolation point as the center of the sphere, determine a spherical region with radius satisfying the following formula: ; updating and saving the distance data set of the sample points to the interpolation points within the spherical region within the spherical region ； ； in, For the spherical region, the first The distance from each sampling point to the interpolation point For the spherical region, the first The coordinates of each sampling point This represents the number of sampling points within the spherical region. S23, set the minimum number of sampling points , determine whether it is greater than or equal to , if the determination result is yes, continue step S24, otherwise, expand the spherical region radius, i.e. let , and return to S22; S24, based on establishing a weight dataset for the sampling points to the interpolation points within the spherical region :​ ； ； wherein is an exponential parameter.

5. The three-dimensional position fingerprint database construction method according to claim 1, characterized by, The step 4 includes combining the sampling point three-dimensional position fingerprint database with the interpolation point three-dimensional position fingerprint database to obtain the final three-dimensional position fingerprint database . 6.An electronic device comprising a memory and a processor, the memory having stored therein a computer program, characterized in that, When the computer program is executed by the processor, the processor implements the method according to any one of claims 1-5.

7. A computer-readable storage medium having stored thereon a computer program, characterized in that, When the computer program is executed by the processor, the method according to any one of claims 1-5 is implemented.

Images