A position information determination method, system, device and storage medium

Abstract

The application discloses a position information determination method, system, device and storage medium, wherein preliminary CSI data is acquired, positioning features meeting a discrimination standard in the preliminary CSI data are extracted, and a respective corresponding CSI data set of each antenna pair is formed; a near neighbor sample set of each antenna pair is determined according to the positioning features on each antenna pair and a preset reference point, and the near neighbor sample set includes a plurality of preset reference points meeting a preset similarity standard in similarity with the positioning features of the antenna pair; frequency statistics of the near neighbor samples in the near neighbor sample sets of all antenna pairs are performed, and position information is determined. The CSI indoor positioning method based on multi-antenna pair fusion fully excavates the CSI features on a plurality of antenna pairs, effectively fuses and statistically selects the near neighbor sample sets of the antenna pairs, and high-precision positioning information is obtained according to the differences in the CSI information on the multi-antenna pairs, thereby improving positioning precision.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of indoor positioning, in particular to a position information determination method, system, device and storage medium. BACKGROUND

[0002] With the development of the Internet of Things and the upgrading of wireless communication equipment, positioning in indoor environments has attracted extensive attention from researchers. Realizing fast and accurate determination of the position information of a target in a positioning scene is a key point in the research of position information application services.

[0003] CSI (Channel State Information) is a channel attribute of the physical layer in wireless communication, which describes the propagation process of a wireless signal from a transmitting end to a receiving end. The existing indoor positioning method based on CSI only considers information on one pair of antennas, ignoring the differences between multiple pairs of antennas, resulting in low positioning accuracy.

[0004] Therefore, how to improve the accuracy of the position information obtained by positioning is a technical problem to be solved by those skilled in the art. SUMMARY

[0005] Based on the above problems, the present application provides a position information determination method, system, device and storage medium to improve the accuracy of positioning.

[0006] To solve the above problems, the technical scheme provided by the embodiments of the present application is as follows:

[0007] The first aspect of the present application provides a position information determination method, comprising:

[0008] obtaining preliminary CSI data, wherein the preliminary CSI data comprises CSI data of each antenna pair in a target range;

[0009] extracting positioning features in the preliminary CSI data that meet a discriminability standard to form a CSI data set corresponding to each antenna pair respectively;

[0010] determining a near neighbor sample set of each antenna pair according to the positioning features on each antenna pair and a preset reference point, wherein the near neighbor sample set comprises a plurality of preset reference points that meet a preset similarity standard with the positioning features of the antenna pair;

[0011] performing frequency statistics on near neighbor samples in the near neighbor sample sets of all antenna pairs to determine position information.

[0012] Optionally, the step of extracting the positioning features in the preliminary CSI data that meet the discriminability standard to form the CSI data set corresponding to each antenna pair respectively comprises:

[0013] According to the prepared CSI data set, an intra-class scatter matrix of each antenna pair and an inter-class scatter matrix of each antenna pair are calculated;

[0014] An optimization function is determined according to the intra-class scatter matrix and the inter-class scatter matrix of each antenna pair, and a positioning feature extraction matrix of each antenna pair is obtained by solving the optimization function;

[0015] According to the positioning feature extraction matrix of each antenna pair, a CSI data set corresponding to each antenna pair is calculated.

[0016] Optionally, the determination of the near neighbor sample set of each antenna pair according to the positioning feature of each antenna pair and the preset reference point comprises:

[0017] A classifier corresponding to each antenna pair is respectively constructed for the positioning feature of each antenna pair;

[0018] The near neighbor sample set of each antenna pair is determined according to the classifier of each antenna pair and the preset reference point.

[0019] Optionally, the determination of the near neighbor sample set of each antenna pair according to the classifier of each antenna pair and the preset reference point comprises:

[0020] According to the CSI data set corresponding to each antenna pair, the mean and the variance of all positioning features at each preset reference point of each antenna pair are calculated;

[0021] A model is constructed according to the mean and the variance of the positioning feature at each preset reference point of each antenna pair, and the model is used to simulate the occurrence of the positioning feature of the antenna pair at each reference point;

[0022] The posterior probability of the occurrence of the positioning feature of each antenna pair at each reference point is calculated, and the reference points meeting a preset probability condition are selected from the calculated posterior probabilities to form a near neighbor sample set.

[0023] Optionally, the frequency statistics of the near neighbor samples in the near neighbor sample set of all antenna pairs to determine the position information comprises:

[0024] The label vector of each near neighbor sample is determined according to the near neighbor samples in the near neighbor sample set of all antenna pairs, and the frequency statistics of all label vectors is performed to obtain the confidence of all near neighbor samples;

[0025] The position information is calculated according to the obtained confidence.

[0026] Optionally, before the prepared CSI data is obtained, the method further comprises:

[0027] The original CSI data of each antenna pair in a target range is obtained;

[0028] Preprocess the original CSI data to obtain preliminary CSI data conforming to a preset data standard.

[0029] Optionally, the preprocessing of the original CSI data comprises:

[0030] Removing outliers in the original CSI data.

[0031] Filtering high-frequency noise in the original CSI data after the outlier removal to obtain the preliminary CSI data.

[0032] The second aspect of the present application provides a position information determination system, comprising:

[0033] A preliminary data acquisition unit configured to acquire preliminary CSI data, the preliminary CSI data comprising CSI data of each antenna pair in a target range;

[0034] A CSI data set determination unit configured to extract positioning features conforming to a discriminability standard from the preliminary CSI data to form a CSI data set corresponding to each antenna pair respectively;

[0035] A near neighbor sample set determination unit configured to determine a near neighbor sample set of each antenna pair according to the positioning features on each antenna pair and preset reference points, the near neighbor sample set comprising a plurality of preset reference points having a similarity conforming to a preset similarity standard with the positioning features of the antenna pair;

[0036] A position information determination unit configured to perform frequency statistics on near neighbor samples in the near neighbor sample sets of all antenna pairs to determine position information.

[0037] The third aspect of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the position information determination method of any one of the first aspect.

[0038] The fourth aspect of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores instructions, and when the instructions run on a terminal device, the terminal device executes the position information determination method of any one of the first aspect.

[0039] Compared with the prior art, the present application has the following beneficial effects:

[0040] By acquiring the preliminary CSI data, extracting positioning features in the preliminary CSI data that meet a discriminant standard, and composing a respective CSI data set for each antenna pair, a near neighbor sample set for each antenna pair is determined according to the positioning features on each antenna pair and preset reference points, the near neighbor sample set including a plurality of preset reference points that have a similarity to the positioning features of the antenna pair meeting a preset similarity standard, and the frequency of the near neighbor samples in the near neighbor sample sets of all antenna pairs is counted to determine the position information. Compared with the prior art that only relies on information on a pair of antenna pairs, the scheme proposed in the present application implements an indoor positioning method based on the fusion of CSI of multiple antenna pairs, fully mines the CSI features on multiple antenna pairs, effectively fuses and statistically selects the near neighbor sample sets of each antenna pair, and thus obtains positioning information with high precision according to the differences in the CSI information on multiple pairs of antenna pairs, thereby improving the positioning precision. BRIEF DESCRIPTION OF DRAWINGS

[0041] To make the technical solutions of the present application or the prior art clearer, the drawings needed in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0042] Figure 1 A flow chart of a position information determination method provided by the present application;

[0043] Figure 2 A schematic diagram of an antenna pair CSI data set determination process provided by the present application;

[0044] Figure 3 A schematic diagram of an antenna pair near neighbor sample set determination process provided by the present application;

[0045] Figure 4 A structural diagram of a position information determination system provided by the present application. DETAILED DESCRIPTION

[0046] To make the technical solutions of the present application or the prior art clearer, the drawings needed in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0047] To make the technical solutions of the present application or the prior art clearer, the drawings needed in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor. To make the technical solutions of the present application or the prior art clearer, the drawings needed in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0048] As described above, with the growing demand for location-based services and applications, positioning in indoor environments has attracted extensive attention from researchers. How to quickly and accurately determine the position information of the target in the positioning scene is the focus of location information application service research. Currently, existing CSI-based indoor positioning methods only consider the information on one pair of antennas, ignoring the differences between multiple pairs of antennas. When selecting different pairs of antennas, the position information generated by positioning may have a large deviation, resulting in low positioning accuracy.

[0049] The method provided by the embodiments of the present application is executed by a background system, for example, can be executed by a positioning background server. The positioning background server can be a server device or a server cluster composed of multiple servers.

[0050] To solve this problem, the embodiments of the present application provide a position information determination method, system, device and storage medium. That is, a CSI device-free indoor positioning method based on multi-antenna pair fusion. This method fully excavates the CSI features of multiple antenna pairs, constructs multiple positioning classifiers based on the Bayesian principle, and effectively fuses them at the near neighbor sample level, avoiding the instability caused by only using the CSI features of one antenna pair, thereby improving the positioning accuracy.

[0051] To facilitate understanding of the position information determination method provided by the embodiments of the present application, the following describes a scenario example of the present application.

[0052] The following describes a position information determination method provided by the present application through an embodiment. Referring to FIG. 1, which is a flowchart of a position information determination method provided by an embodiment of the present application. The execution subject of the method flow is a server, and further, the subject can be a positioning system in the server. The method includes the following steps: Figure 1

[0053] S101: Obtain preliminary CSI data.

[0054] The preliminary CSI data includes the CSI data of each antenna pair in the target range, which is used for subsequent feature extraction. The target range can be a pre-planned indoor environment.

[0055] In a possible implementation, before the step of obtaining preliminary CSI data, the method further includes steps A1-A2:

[0056] A1: Obtain the original CSI data of each antenna pair in the target range.

[0057] ​In the actual application scenario, the collection device can be a TP-LINK router and a Dell desktop computer with an InterWIFI Link 5300 wireless network card. The experimental target stands at the planned position in turn, keeps the body still, and another person operates the desktop computer to collect the CSI data. Since the router used in the experiment has only two antennas and the 5300 network card has three antennas, there are a total of six pairs of antenna pairs, and the collected CSI data format is 2*3*30.

[0058] A2: Preprocessing the original CSI data to obtain preliminary CSI data conforming to a preset numerical standard.

[0059] The preset numerical standard is used to filter out redundant data that is not available in the original CSI data. For example, a filtering standard for abnormal values in the original CSI data can be set, and / or a filtering standard for high-frequency noise in the original CSI data can be set. The above two filtering standards are examples and do not limit the preprocessing. The standard can be adaptively adjusted or increased or decreased according to actual needs.

[0060] In one possible implementation, step A2 preprocesses the original CSI data, including steps A201-A202:

[0061] A201: Remove abnormal values in the original CSI data.

[0062] Since the original CSI data collected by the device contains not only useful information but also abnormal values caused by interference due to instruments and uncertain factors. In order to remove abnormal values, a Hampel filter can be used, which can identify abnormal values and replace them with more representative values.

[0063] A202: Filter out high-frequency noise in the original CSI data after abnormal value processing to obtain preliminary CSI data.

[0064] Since noise caused by changes in the surrounding environment and electromagnetic interference in an indoor environment causes many burrs in the CSI data after abnormal value processing, these burrs are caused by high-frequency noise. In the actual application scenario, in order to filter out high-frequency noise, a wavelet threshold shrinkage method can be used, which can effectively remove random fluctuations in the CSI signal caused by noise.

[0065] It should be noted that the above processing methods for abnormal values and noise are only examples and do not limit the processing method in this step. In the actual application scenario, the processing method can be selected and adaptively adjusted according to actual needs.

[0066] S102: Extract positioning features in the preliminary CSI data that meet the discriminability standard to form a respective CSI data set for each antenna pair.

[0067] The preliminary CSI data obtained in the foregoing step is analyzed to extract positioning features with the highest discriminability for each antenna pair. The discriminability standard for determining whether the positioning features of each antenna pair have discriminability can be set according to actual requirements.

[0068] In a possible implementation, the step of extracting positioning features in the preliminary CSI data that meet the discriminability standard to form a respective CSI data set for each antenna pair includes steps B1-B3, which can be seen in Figure 2 , Figure 2 A schematic diagram of an antenna pair CSI data set determination process provided by an embodiment of the present application.

[0069] In an actual application scenario, the linear discriminant analysis can be used to extract positioning features with the highest discriminability for each antenna pair from the preliminary CSI data. An optimization function is constructed, and the Lagrange multiplier method is used to obtain a positioning feature extraction matrix for each antenna pair to generate positioning features for each antenna pair.

[0070] It should be noted that only one antenna pair is described in the following steps, and the s-th antenna pair is used to represent one of the several antenna pairs in the present solution. In an actual application scenario, each antenna pair can be calculated simultaneously, and the following example description of the s-th antenna pair does not limit the number of antenna pairs in the present application.

[0071] B1: Calculate the within-class scatter matrix of each antenna pair and the between-class scatter matrix of each antenna pair according to the preliminary CSI data set.

[0072] Taking the s-th antenna pair as an example, the within-class scatter matrix of the s-th antenna pair and the between-class scatter matrix of the s-th antenna pair are calculated according to the CSI data set of the s-th antenna pair after data preprocessing.

[0073] Suppose the CSI data set of the s-th antenna pair after data preprocessing at the j-th reference point can be represented as F sj , where F sj =(csi sj1 ,csi sj2 ,csi sj3 ,…,csi sjm ), and m represents the number of CSI data packets at each reference point. Then, the CSI data set of the s-th antenna pair after data preprocessing at n reference points can be represented as F s =[F s1 ,F s2 ,Fs3 ,…,F sn ] T , whose expansion is

[0074]

[0075] Because the 5300 network card can collect the CSI data on 30 subcarriers, each CSI data packet is a 1 x 30 row vector, and in formula (1), the CSI amplitude data set F s after data preprocessing is an h x 30 matrix, where h = n x m. Linear discriminant analysis maps the CSI data set with n classification labels to a feature space, which is represented by t basis vectors, and the positioning feature extraction matrix V s on the s-th antenna pair is s (v s1 ,v s2 ,…,v st ). To obtain the positioning feature extraction matrix V s , the within-class scatter matrix Φ s on the s-th antenna pair and the between-class scatter matrix Ψ s on the s-th antenna pair need to be calculated, and the expressions are as follows:

[0076]

[0077] In formula (3), csi sjp represents the p-th CSI data value in the j-th class of the s-th antenna pair, and represents the average value of the CSI data in the j-th class of the s-th antenna pair. The formula is as follows:

[0078]

[0079] Similarly, the between-class scatter matrix Ψ s on the s-th antenna pair can be obtained as follows:

[0080]

[0081] In formula (4), the between-class scatter matrix Ψ s represents the sum of the distances from the mean value of each class in the s-th antenna pair to the mean value of all classes. In formula (5), csi represents the average value of the CSI data in the j-th class of the s-th antenna pair, and represents the average value of all CSI data in the s-th antenna pair, and the calculation formula is as follows:

[0082]

[0083] B2: determining an optimization function according to the intra-class scatter matrix and the inter-class scatter matrix of each antenna pair, solving the optimization function to obtain a positioning feature extraction matrix of each antenna pair.

[0084] In an actual application scenario, a target function can be constructed according to the Fisher discriminant criterion, and is converted into an optimization function through the property of the generalized Rayleigh quotient. The method for constructing the target function can be the Fisher discriminant criterion, and the method can be adaptively adjusted according to actual requirements in the application process.

[0085] It can be known from the Fisher discriminant criterion that when the ratio of the inter-class distance of different class samples to the intra-class distance of the same class samples is maximum, the basis vector of the feature space can be obtained, and therefore the following target function can be constructed.

[0086]

[0087] Equation (6) represents the generalized Rayleigh quotient of Ψ s and Φ s , and maximizing the generalized Rayleigh quotient can be converted into the following optimization function:

[0088]

[0089] The optimization function is solved by using the Lagrange multiplier method, and the positioning feature extraction matrix of the s-th antenna pair is determined.

[0090] By using the Lagrange multiplier method, equation (7) can be equivalent to the following equation (8):

[0091] J(V s )=(V s ) T Ψ s V s -λ((V s ) T Φ s V s -1) (8)

[0092] In equation (8), λ is the Lagrange multiplier. The s-th antenna pair positioning feature extraction matrix V s is derived, and the following equation (11) can be obtained:

[0093]

[0094]

[0095] (Φ s ) -1 Ψ s V s =λV s (11)

[0096] From formula (11), the positioning feature extraction matrix V s of the s-th antenna pair is composed of the eigenvectors of the matrix (Φ s ) -1 Ψ s . First, the eigenvalues and eigenvectors of the matrix (Φ s ) -1 Ψ s are calculated, the eigenvalues are sorted in descending order, and then the eigenvectors corresponding to the first t eigenvalues are selected to form the positioning feature extraction matrix V s of the s-th antenna pair.

[0097] B3: According to each antenna pair positioning feature extraction matrix, the respective CSI data set corresponding to each antenna pair is calculated.

[0098] According to the s-th antenna pair positioning feature extraction matrix, the CSI data set composed of the positioning features on the s-th antenna pair is calculated.

[0099] Since the value of the positioning feature extraction matrix V s is known, the CSI data set can be changed to:

[0100] G s = F s V s (12)

[0101] In formula (12), G s represents the CSI data set composed of the positioning features on the s-th antenna pair, which is an hxt matrix, F s represents the preprocessed CSI data set on the s-th antenna pair, which is an h×30 matrix, and V s is the positioning feature extraction matrix on the s-th antenna pair, which can be represented as a 30xt matrix.

[0102] S103: Determine the near neighbor sample set of each antenna pair according to the positioning features on each antenna pair and the preset reference points.

[0103] The near neighbor sample set includes a plurality of preset reference points that meet the preset similarity standard with the positioning features of the antenna pair. By comparing the relationship between the positioning features on each antenna pair and each preset reference point, a plurality of reference points that meet the requirements of the near neighbor sample set are obtained, and the near neighbor sample set of the antenna pair is generated. It should be noted that the near neighbor sample set is for a certain antenna pair, and the reference points that meet the requirements are compared with the positioning features of the antenna pair, and have nothing to do with other antenna pairs being calculated at the same time. The near neighbor sample sets obtained by each antenna pair are relatively independent.

[0104] In a possible implementation, the step C1-C2 of determining the near neighbor sample set of each antenna pair according to the positioning features on each antenna pair and the preset reference points comprises the following steps:

[0105] C1: respectively constructing a classifier corresponding to each antenna pair for the positioning features on each antenna pair.

[0106] C2: determining the near neighbor sample set of each antenna pair according to the classifier of each antenna pair and the preset reference points.

[0107] For the positioning features on each antenna pair generated in step S102, a classifier is respectively constructed, and the output of each classifier is K near neighbor samples of each antenna pair that meet a similarity criterion with the fingerprint database. The similarity criterion and the number of samples in the near neighbor sample set can be set according to actual requirements. The classifier can be a classifier constructed based on the Bayesian principle.

[0108] In a possible implementation, the step C2 of determining the near neighbor sample set of each antenna pair according to the classifier of each antenna pair and the preset reference points comprises the following steps C201-C203, which can be referred to as Figure 3 , Figure 3 The process diagram of determining the near neighbor sample set of an antenna pair provided by the embodiment of the application is shown in FIG. 2.

[0109] C201: calculating the mean and variance of all positioning features at each preset reference point for each antenna pair according to the CSI data set corresponding to each antenna pair.

[0110] According to the CSI data set composed of the positioning features on the s-th antenna pair, the mean and variance of all positioning features at each reference point of the s-th antenna pair are calculated.

[0111] Suppose that after linear discriminant analysis, each CSI data packet is left with t positioning features, and the m-th CSI data packet at the j-th reference point of the s-th antenna pair can be expressed as The CSI data set after positioning feature extraction at the j-th reference point of the s-th antenna pair can be expressed as G sj , where G sj = [g sj1 ,g sj2 ,g sj3 ,…,g sjm ], where m represents the number of CSI data packets at each reference point. The mean of each positioning feature in the CSI data at the j-th reference point of the s-th antenna pair is as follows:

[0112]

[0113] wherein, in formula (13), the mean of each positioning feature in the CSI data at the j-th reference point of the s-th antenna pair is as follows: The mean value of the first feature at the jth reference point of the st antenna pair is represented. The CSI value of the first feature of the pth data packet at the jth reference point of the st antenna pair is represented. Similarly, according to the above content, the mean value of all features at the jth reference point of the st antenna pair can be obtained

[0114] The variance of each feature in the CSI data at the jth reference point of the st antenna pair is shown as follows:

[0115]

[0116] Similarly, according to the above formula (14), the variance of all features at the jth reference point of the st antenna pair can be obtained

[0117] C202: Constructing a model according to the mean value and variance of the positioning feature at each preset reference point of each antenna pair.

[0118] The model is used to simulate the occurrence of the positioning feature of the antenna pair at each reference point; the multivariate Gaussian distribution model of the jth reference point of the st antenna pair is:

[0119]

[0120] In formula (15), the product of the mean value of all features at the jth reference point of the st antenna pair is represented. The mean value of all features at the jth reference point of the st antenna pair is represented, and in the formula The product of the variance of all features at the jth reference point of the st antenna pair is represented, and similarly, the multivariate Gaussian distribution model of all reference points of the st antenna pair can be calculated, and after normalization, the conditional probability P(g si | l si ) that the CSI sample g j of the st antenna pair occurs at the jth reference point can be calculated.

[0121]

[0122] In formula (16), P(g si | l j ) represents the conditional probability P(g si | l si ) that the CSI sample g j of the st antenna pair occurs at the jth reference point.

[0123] C203: Calculating the posterior probability of the occurrence of the positioning feature of each antenna pair at each reference point, and selecting the reference points meeting the preset probability condition from the calculated several posterior probabilities to form a near neighbor sample set.​

[0124] In an actual application scenario, the posterior probability of the positioning feature of the s th antenna pair appearing at each reference point can be obtained by using the Bayesian principle, the reference points corresponding to the top K probability values are selected in descending order, and the K nearest neighbor samples are obtained. The determination method of the posterior probability can be the Bayesian principle, and the application process can be adjusted according to actual needs.

[0125] According to the Bayesian principle, the posterior probability P (l si |g j ) of the CSI sample g si of the s th antenna pair appearing at the j th reference point can be calculated, and the expression is as follows:

[0126]

[0127] In formula (17), P (l j ) represents the probability of the j th reference point appearing in all reference points, which is assumed to be uniformly distributed here. P (g si |l j ) represents the conditional probability of the CSI sample g si of the s th antenna pair appearing at the j th reference point. Similarly, the posterior probability of the CSI sample g si of the s th antenna pair appearing at all reference points can be obtained, and then the reference points corresponding to the top K probability values are selected in descending order, and the K nearest neighbor samples are obtained.

[0128] S104: Count the frequency of the nearest neighbor samples in the nearest neighbor sample set of all antenna pairs, and determine the position coordinates.

[0129] In one possible implementation, the step of counting the frequency of the nearest neighbor samples in the nearest neighbor sample set of all antenna pairs and determining the position coordinates includes steps D1-D2:

[0130] That is, the K nearest neighbor samples most similar to the positioning feature of each antenna pair in step S103 are uniformly counted to obtain the confidence of all nearest neighbor samples, and finally the estimated position coordinates are obtained by using the confidence regression method.

[0131] D1: Determine the label vector of each nearest neighbor sample according to the nearest neighbor samples in the nearest neighbor sample set of all antenna pairs, count the frequency of all label vectors, and obtain the confidence of all nearest neighbor samples.

[0132] In this application, 6K nearest neighbor samples are obtained by collecting 6 antenna pairs. Specifically, it is assumed that q iThe label of a neighbor sample of a certain Bayesian-based classifier, which can be represented as (x i ,y i ). Since 6 Bayesian-based classifiers are constructed, 6K neighbor samples can be obtained, and the label vector of all classifiers is q = [q1, q2, …, q 6K ], and then frequency statistics is performed to obtain the confidence of all neighbor samples. The confidence of a certain neighbor sample q j is represented as follows:

[0133]

[0134] In formula (18), q j represents a certain neighbor sample, P(q j ) represents the confidence of a certain neighbor sample q j , which is a probability value. q i represents the label of a neighbor sample of a certain Bayesian-based classifier. q j == q i represents comparison of the neighbor sample q j and the neighbor sample label q i , to determine whether the sample and the sample label are equal, and the confidence of the neighbor sample is obtained according to the comparison.

[0135] D2: calculating the position coordinates according to the obtained confidence.

[0136] In the confidence regression manner, the final estimated position coordinates are obtained, and the expression is as follows:

[0137]

[0138] In formula (19), L j represents the position coordinates of the jth reference point, which can be represented as (x j ,y j ), and the position coordinate set can be represented as L = [l1, l2, …, l n ]. represents the estimated position coordinates, which can be represented as

[0139] The above are some specific implementation manners of the position information determination method provided by the embodiments of the present application. Based on this, the present application further provides a corresponding system for position information determination. The system provided by the embodiments of the present application will be introduced from the perspective of functional modularization. Figure 4 A position information determination system structure diagram provided by an embodiment of the present application.

[0140] The system comprises:

[0141] The preparation data acquisition unit 201 is configured to acquire preparation CSI data, wherein the preparation CSI data comprises CSI data of each antenna pair within a target range;

[0142] The CSI data set determination unit 202 is configured to extract positioning features in the preparation CSI data that meet a discriminant criterion, and to form a CSI data set corresponding to each antenna pair respectively.

[0143] The near neighbor sample set determination unit 203 is configured to determine a near neighbor sample set of each antenna pair according to the positioning features on each antenna pair and preset reference points, wherein the near neighbor sample set comprises a plurality of preset reference points that meet a preset similarity criterion with the positioning features of the antenna pair.

[0144] The position information determination unit 204 is configured to perform frequency statistics on the near neighbor samples in the near neighbor sample sets of all antenna pairs, and to determine position information.

[0145] Optionally, the CSI data set determination unit is specifically configured to calculate an intra-class scatter matrix of each antenna pair and an inter-class scatter matrix of each antenna pair according to the preparation CSI data set; determine an optimization function according to the intra-class scatter matrix and the inter-class scatter matrix of each antenna pair, and solve the optimization function to obtain a positioning feature extraction matrix of each antenna pair; and calculate a CSI data set corresponding to each antenna pair according to the positioning feature extraction matrix of each antenna pair.

[0146] Optionally, the near neighbor sample set determination unit comprises a classifier construction unit and a determination execution unit.

[0147] The classifier construction unit is configured to construct a classifier corresponding to each antenna pair respectively for the positioning features on each antenna pair.

[0148] The determination execution unit is configured to determine a near neighbor sample set of each antenna pair according to the classifier of each antenna pair and the preset reference points.

[0149] Optionally, the determination execution unit is specifically configured to calculate a mean and a variance of all positioning features at each preset reference point for each antenna pair according to the CSI data set corresponding to each antenna pair; construct a model according to the mean and the variance of the positioning features at each preset reference point for each antenna pair, wherein the model is used to simulate the occurrence of the antenna pair positioning features at each reference point; calculate a posterior probability of the positioning features of each antenna pair occurring at each reference point, and select reference points that meet a preset probability condition from the calculated plurality of posterior probabilities to form a near neighbor sample set.

[0150] Optionally, the position coordinate determination unit is configured to determine a label vector of each neighbor sample according to neighbor samples in the neighbor sample set of each antenna pair, count frequencies of all label vectors to obtain confidence of all neighbor samples, and calculate the position coordinate according to the obtained confidence.

[0151] Optionally, the system further comprises,

[0152] a raw CSI data acquisition unit configured to acquire raw CSI data of each antenna pair in a target range;

[0153] a preprocessing unit configured to preprocess the raw CSI data to obtain preliminary CSI data meeting a preset standard.

[0154] The preprocessing unit is specifically configured to remove outliers in the raw CSI data, and filter high-frequency noise in the raw CSI data after the outlier removal to obtain the preliminary CSI data.

[0155] Embodiments of the present application further provide a corresponding device and a computer storage medium for implementing the position information determination method provided by the embodiments of the present application.

[0156] The device comprises a memory and a processor, the memory is configured to store instructions or codes, and the processor is configured to execute the instructions or codes to enable the device to perform the position information determination method according to any one of the embodiments of the present application.

[0157] The computer storage medium stores codes, and when the codes are executed, a device running the codes implements the position information determination method according to any one of the embodiments of the present application.

[0158] It should be noted that the embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts of each embodiment can be referred to each other. For the system or device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant part can be referred to the method part.

[0159] It should be understood that, in the application, "at least one" means one or more, "multiple" means two or more. "And / or" is used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, "A and / or B" can mean: only A, only B, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after it. "At least one of the following" or similar expressions means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c, can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0160] It should also be noted that, in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0161] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of the two. The software module can be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0162] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, the application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method of position information determination, characterized by, The method comprises the following steps: obtaining preliminary CSI data, wherein the preliminary CSI data comprises CSI data of each antenna pair within a target range; extracting positioning features in the preliminary CSI data that meet a discriminability standard to form a CSI data set corresponding to each antenna pair respectively; determining a near neighbor sample set of each antenna pair according to the positioning features on each antenna pair and preset reference points, wherein the near neighbor sample set comprises a plurality of preset reference points that meet a preset similarity standard with the positioning features of the antenna pair; counting the frequency of near neighbor samples in the near neighbor sample sets of all antenna pairs to determine position information. The step of extracting the positioning features in the preliminary CSI data that meet the discriminability standard to form the CSI data set corresponding to each antenna pair respectively comprises the following steps: calculating an intra-class scatter matrix of each antenna pair and an inter-class scatter matrix of each antenna pair according to the preliminary CSI data; determining an optimization function according to the intra-class scatter matrix and the inter-class scatter matrix of each antenna pair, and solving the optimization function to obtain a positioning feature extraction matrix of each antenna pair; calculating the CSI data set corresponding to each antenna pair respectively according to the positioning feature extraction matrix of each antenna pair.

2. The method of claim 1, wherein, The step of determining the near neighbor sample set of each antenna pair according to the positioning features on each antenna pair and the preset reference points comprises the following steps: constructing a classifier corresponding to each antenna pair respectively for the positioning features on each antenna pair; determining the near neighbor sample set of each antenna pair according to the classifier of each antenna pair and the preset reference points.

3. The method of claim 2, wherein, The step of determining the near neighbor sample set of each antenna pair according to the classifier of each antenna pair and the preset reference points comprises the following steps: calculating the mean and variance of all positioning features at each preset reference point for each antenna pair according to the CSI data set corresponding to each antenna pair respectively; constructing a model according to the mean and variance of the positioning features at each preset reference point for each antenna pair, wherein the model is used to simulate the occurrence of the positioning features of the antenna pair at each reference point; calculating the posterior probability of the positioning features of each antenna pair occurring at each reference point, and selecting reference points that meet a preset probability condition from the calculated posterior probabilities to form the near neighbor sample set.

4. The method of claim 1, wherein, The step of counting the frequency of the near neighbor samples in the near neighbor sample sets of all antenna pairs to determine the position information comprises the following steps: determining a label vector of each near neighbor sample according to the near neighbor samples in the near neighbor sample sets of all antenna pairs, counting the frequency of all label vectors to obtain the confidence of all near neighbor samples, and calculating the position information according to the confidence obtained by the counting. Before the step of obtaining the preliminary CSI data, the method further comprises the following steps:

5. The method of claim 1, wherein, obtaining original CSI data of each antenna pair within a target range; preprocessing the original CSI data to obtain preliminary CSI data that meets a preset data standard. The step of preprocessing the original CSI data comprises the following steps:

6. The method of claim 5, wherein, removing outliers in the original CSI data; filtering high-frequency noise in the original CSI data after the outlier processing to obtain the preliminary CSI data. The system comprises:

7. A position information determination system, characterized by a preliminary data acquisition unit configured to obtain preliminary CSI data, wherein the preliminary CSI data comprises CSI data of each antenna pair within a target range. ​ The CSI dataset determining unit is configured to extract positioning features meeting a discriminant criterion from the preliminary CSI data, and to form a respective CSI dataset for each antenna pair; The near neighbor sample set determining unit is configured to determine a near neighbor sample set for each antenna pair according to the positioning features on each antenna pair and preset reference points, the near neighbor sample set including a plurality of preset reference points meeting a preset similarity criterion with the positioning features of the antenna pair; The position information determining unit is configured to perform frequency statistics on the near neighbor samples in the near neighbor sample sets of all antenna pairs, and to determine position information. The CSI dataset determining unit is specifically configured to calculate an intra-class scatter matrix of each antenna pair and an inter-class scatter matrix of each antenna pair according to the preliminary CSI data; determine an optimization function according to the intra-class scatter matrix and the inter-class scatter matrix of each antenna pair, and solve the optimization function to obtain a positioning feature extraction matrix of each antenna pair; and calculate a respective CSI dataset for each antenna pair according to the positioning feature extraction matrix of each antenna pair.

8. An electronic device, comprising: The computer readable storage medium stores instructions, and when the instructions run on the terminal device, the terminal device performs the position information determining method according to any one of claims 1-6. The computer readable storage medium stores instructions, and when the instructions run on the terminal device, the terminal device performs the position information determining method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, ​

Images