CN104968004B - Indoor WLAN fingerprint locations access point deployment method based on user location secret protection - Google Patents

Abstract

Indoor WLAN fingerprint locations access point deployment method of the present invention based on user location secret protection; first with improved anonymity degree (Anonymity Degree; AD), without validity (Ineffectiveness Degree; ID user location privacy and LBS service quality are portrayed) respectively; secondly according to stream of people's distribution situation different in different AP deployment way and localization region; calculate anonymity degree and without validity; it finally assigns anonymity degree and builds optimization object function without the certain weight of validity, and pass through to search for and find optimal AP deployment way.This method protects the location privacy of user while LBS service quality is ensured.

Description

Deployment method of indoor WLAN fingerprint positioning access point based on user location privacy protection

technical field

The invention belongs to radio communication technology, and in particular relates to an indoor WLAN fingerprint positioning access point deployment method based on user location privacy protection.

Background technique

Location-based service (Location Based Service, LBS) refers to an application service that obtains the user's location information through the mobile network and positioning technology and provides the service information related to the location to the user. Through LBS, the user can know where he is now. Where, at the same time, various information related to life, commerce and transportation can be queried. Therefore, LBS is considered to be one of the "killer" services in mobile computing, and has good market value and development prospects.

The composition of LBS mainly includes mobile terminals (such as mobile phones, tablet computers, etc.), positioning systems (used to determine the location information of users), mobile communication networks and service content providers. At present, the more popular outdoor wireless positioning system is the Global Positioning System (GPS), which estimates the terminal position by capturing GPS satellite signals and measuring the arrival delay of at least 4 GPS satellite signals in orbit. This system can provide approximately global coverage. Range, high precision, and all-weather continuous positioning and navigation capabilities, but in complex indoor environments, satellite signals will decline sharply, and positioning performance is not ideal. Common indoor wireless positioning systems include Bluetooth positioning systems, radio frequency identification (RFID) positioning systems, ZigBee positioning systems and WLAN positioning systems. Due to the continuous popularization of WLAN, the positioning technology based on WLAN has been greatly developed. Currently, mainstream WLAN positioning technologies can be classified into the following four categories: time of arrival (or time difference) positioning, angle of arrival (or angle difference) positioning, propagation model positioning, and location fingerprint positioning. Among them, location fingerprint positioning has been widely used because of its high positioning accuracy and no need to add additional hardware devices.

The system flow of LBS is usually as follows: First, the mobile user makes a service request to the LBS server, and at the same time reports the location information obtained by the positioning system and the query content to the LBS server; then, the server processes the user's request according to the reported location information; finally, The server sends the service content to the mobile user. Since the user's location information and query content are easily obtained by third-party attackers, users need to provide their own location information in order to obtain LBS, and the quality of LBS obtained by users is directly proportional to the accuracy of the location information reported by users. Therefore, , how to deal with the contradiction between LBS quality and user location privacy protection reasonably and effectively is one of the key issues of user privacy protection in LBS.

Location privacy is one of the important components of user privacy. At present, most studies mainly consider the location privacy protection when the user reports location information to the LBS server in the process of the LBS system, such as the area coverage privacy protection technology. The location information of the user is blurred from a point to a spatial area. At this time, even if the attacker obtains the location area information reported by the user to the LBS server, the exact location of the user cannot be obtained. However, in the LBS based on the indoor WLAN positioning system, even if the area coverage privacy protection technology is adopted, if the user's RSS fingerprint and fingerprint database in the positioning stage are leaked at the same time, the attacker can use a certain positioning algorithm (such as the KNN algorithm) to narrow down where the user is located. Therefore, the risk of user location privacy exposure will increase. Based on this, this paper aims to solve the location privacy protection problem in the case where the area information reported by the user to the LBS server, the user's RSS fingerprint, and the fingerprint database are all leaked when the area coverage privacy protection technology is used in the LBS based on the indoor WLAN fingerprint positioning system. . When both the user's RSS fingerprint and the fingerprint database are leaked, the attacker can estimate the user's location through a positioning algorithm (such as the KNN algorithm), that is, find the user's location estimation point. Considering that there is a certain positioning error in the WLAN positioning system, the user's The location information can be narrowed down to a circle with the estimated user position as the center and the positioning error as the radius. If there are K users in the circle, the user's location information can be further narrowed down to one of the K user locations. It can be seen that if the positioning accuracy of the system is very high, although the user's LBS quality can be guaranteed, the user's location privacy is also very easy to leak at this time, and when using the area coverage privacy protection technology, as long as the user's location estimation point is reported In the area, the quality of LBS can be effectively guaranteed. Based on this, the present invention proposes a new AP deployment method, which can effectively guarantee the quality of LBS while protecting user location privacy.

Contents of the invention

The purpose of the present invention is to provide an indoor WLAN fingerprint positioning access point deployment method based on user location privacy protection. On the premise of ensuring the quality of LBS, combined with different crowd distribution in the target area, the protection of user location privacy can be realized. And the rapid optimization of AP deployment methods.

The indoor WLAN fingerprint positioning access point deployment method based on user location privacy protection according to the present invention comprises the following steps:

Step 1. Set the weight coefficient μ, the radius R of the location area reported by the target user to the LBS server, and the unit of R is meters;

Step 2, initialization, let i=1, E_current=0, E_best=0, wherein, i is the count amount, E_current is the objective function value corresponding to the current solution when used to store the optimization search, and E_best is the best value when used to store the optimization search The objective function value corresponding to the optimal solution;

Step 3. According to the number of APs, a new AP deployment mode is generated by disturbance, and the AP candidate position label number corresponding to the newly generated AP deployment mode is stored in the matrix slo_new;

Step 4. Use the KNN (K-Nearest Neighbor) algorithm to calculate the position estimation points and corresponding positioning errors of each test point under the current AP deployment mode, and store them in the matrix Location and errors respectively;

Step 5. All elements in the matrix errors are rounded up and stored in the matrix d;

Step 6, make j=1, where j is the count of the number of test points, assuming that the test point is the location point of the user who needs LBS;

Step 7. Determine whether the value of d(j,1) is less than R, that is, whether the positioning error of the test point after rounding up is smaller than the reported location area radius R, where d(j,1) is the positioning error at test point j The value after rounding up; if yes, go to step 8; if not, go to step 15;

Step 8, make k=d(j,1); Wherein k is counting quantity;

Step 9: Take the estimated point coordinates Location(j,:) of the jth test point as the center of the circle, k as the radius, count the number of users in the circle, calculate the Euclidean distance from different users to the center of the circle, and store it in the matrix Distance ;

Step ten, assuming that there are U _k users in the circular domain, then calculating the average information entropy selected as the target user in the circular domain with a radius of k, and storing it in the matrix H (k, 1);

Step 11. When the target user is at the test point j, the attacker selects the anonymity degree Ad(k,1) of the target user in a circle with a radius of k, which reflects the privacy degree of the user;

Step 12, superimposing Ad(k,1) under different radius k into the matrix element AD(j,1), that is, AD(j,1)=AD(j,1)+Ad(k,1), Among them, AD(j, 1) is used to store the total anonymity of the target user selected under different radius k when the target user is at the jth test point;

Step 13, let k=k+1, where k is equal to the radius of the attacker looking for the target user, and calculate the degree of anonymity corresponding to each value of k;

Step 14, determine whether k is less than or equal to R, if so, proceed to step 9; if not, proceed to step 16;

Step 15. Define the test points with positioning errors greater than R as out-of-boundary points, and set the number of out-of-boundary points to be r, where r is the number of d(j, 1) (j=1,...,Num_T) values greater than R ;

Step sixteen, let j=j+1;

Step seventeen, judge whether j is less than Num_T, wherein, Num_T is the total number of test points; if so, then enter step seven; if not, then enter step eighteen;

Step 18, calculate the average anonymity degree of the user under the current AP deployment mode;

Step 19, calculating the average invalidity degree of users under the current AP deployment mode;

Step 20, calculating the objective function value f, where f is the objective function value under the new AP deployment mode generated by the disturbance;

Step 21. Determine whether f is greater than E_current; if so, proceed to step 22; if not, proceed to step 25;

Step 22, set slo_current=slo_new; E_current=f, wherein slo_current is used to store the current AP deployment method during the optimization search; slo_new is used to store the new AP deployment method obtained due to the optimization search disturbance;

Step 23, judging whether f is greater than E_best; if so, proceed to step 24; if not, proceed to step 25;

Step 24, set slo_best=slo_new, E_best=f; wherein, slo_best is the optimal AP deployment method for storage optimization search;

Step 25, let i=i+1;

Step 26. Determine whether i is less than If yes, proceed to step 3; if not, proceed to step 27, where Num_AP is the total number of APs, and AP_candidate is the number of AP candidate positions; Indicates the number of combinations for selecting Num_AP different AP positions from the AP_candidate AP candidate positions in a permutation and combination manner;

Step 27, output slo_best.

The matrix d is:

in, Indicates rounding up.

In the tenth step, the calculation formula of the average information entropy is:

The calculation formula of information entropy H(u) is:

Among them, U _k is the number of people contained in the circle of radius k; p(u) is the probability that the uth person in the circle of radius k is the user; H(u) is the probability of the uth person in the circle of radius k User's information entropy; Distance(u,1) is the Euclidean distance from the uth person in a circle of radius k to the center of the circle.

In the eleventh step, the formula for calculating the degree of anonymity Ad(k, 1) of the target user is:

Ad(k,1)=2 ^H(k,1) .

In the eighteenth step, the formula for calculating the average degree of anonymity of the user under the current AP deployment mode is:

Among them, Num_T is the total number of test points.

In the nineteenth step, the formula for calculating the average invalidity degree of the user under the current AP deployment mode is:

Among them, r is the number of out-of-boundary points; Num_T is the total number of test points.

In said step 20, the calculation formula of said objective function value f is:

f=μ·Aver_AD+(1-μ)·Aver_ID

Among them, f is the objective function value under the new AP deployment mode generated by the disturbance; μ is the weight coefficient.

The present invention has the following advantages: the method is mainly aimed at the LBS based on the indoor WLAN fingerprint positioning system, the area information provided by the user to the LBS server, the user received signal strength (Received Signal Strength, RSS) fingerprint and In the high-risk situation where the fingerprint database is leaked, a new indoor wireless local area network (Wireless Local Area Network, WLAN) fingerprint positioning access point (Access Point, AP) deployment method is proposed to protect the user's location privacy. This method first uses the improved anonymity and ineffectiveness to characterize the user's location privacy and LBS quality respectively. Secondly, according to different AP deployment methods and different crowd distribution in the target area, the anonymity and ineffectiveness are calculated. Finally, through Different weights are given to the degree of anonymity and ineffectiveness to construct the optimization objective function, and at the same time obtain the optimal AP deployment method. It can be seen that this method protects the user's location privacy while ensuring the quality of LBS. The invention can be applied to indoor radio communication network environment.

Description of drawings

Fig. 1 a is the flowchart of step 1 to step 12 in the present invention;

Fig. 1b is the flowchart of step 13 to step 27 in the present invention;

Figure 2 shows a schematic diagram of calculating the degree of anonymity that characterizes user privacy;

3 is a schematic diagram of the real environment of the present invention, which is divided into two areas. There are 38 reference points (marked as "·") in area 1 and 15 reference points (marked as "·") in area 2. In the figure, the five-pointed star symbol represents the origin of the coordinates. The horizontal and vertical coordinate distances of adjacent reference points in area 1 are 2m and 1.2m respectively, and the horizontal and vertical coordinate distances of adjacent reference points in area 2 are 1.2m and 3m respectively. There are 10 AP candidate positions are marked as 1-10 respectively;

Fig. 4 is the crowd distribution and target user position randomly generated in the experimental environment of the present invention;

Figures 5a, 5b and 5c respectively show that in the real experimental environment, when the weight coefficient u=0.7 and the reporting radius R=10m, the weighting coefficient u=0.3 and the reporting radius R=10m, the weighting coefficient u=0.5 and the reporting radius R = 5m, the anonymity comparison results of the indoor WLAN fingerprint positioning AP deployment method based on user location privacy protection proposed by the present invention and the traditional AP deployment method based on positioning error minimization and random selection AP deployment method, wherein, based on positioning error In the minimized AP deployment method, the AP deployment method corresponding to the minimum average positioning error of all test points is taken as the optimal AP deployment method, while in the randomly selected AP deployment method, the AP deployment method generated by random selection is used as the optimal AP deployment method;

Figures 6a, 6b and 6c respectively show that in the real experimental environment, when the weight coefficient u=0.7 and the reporting radius R=10m, the weighting coefficient u=0.3 and the reporting radius R=10m, the weighting coefficient u=0.5 and the reporting radius R When =5m, the indoor WLAN fingerprint positioning AP deployment method based on user location privacy protection proposed by the present invention and the traditional AP deployment method based on positioning error minimization and the invalid degree comparison result of randomly selecting the AP deployment method;

Figures 7a, 7b, 7c, 7d, 7e, 7f, 7g, 7h and 7i respectively show the real experimental environment, when the number of APs is 1, 2, 3, 4, 5, 6, 7, 8 and 9 , the indoor WLAN fingerprint positioning AP deployment method based on user location privacy protection proposed by the present invention, the traditional AP deployment method based on positioning error minimization and the random selection AP deployment method report the user location estimation point within the radius R = 10m Error cumulative distribution comparison results. Since the research background of the present invention is aimed at the LBS of the indoor WLAN fingerprint positioning system, the area information provided by the user to the LBS server, the user's received signal strength fingerprint, and the fingerprint database are all leaked. In this case, the attacker only needs to match the fingerprint data within the reported radius R. Therefore, if the error of the user's location estimation point is larger, the user's location privacy protection is better. Based on this, the indoor WLAN fingerprint positioning AP deployment method based on user location privacy protection proposed by the present invention has better user location privacy protection;

Figures 8a, 8b, 8c, 8d, 8e, 8f, 8g, 8h and 8i respectively show the real experimental environment, when the number of APs is 1, 2, 3, 4, 5, 6, 7, 8 and 9 When the indoor WLAN fingerprint positioning AP deployment method based on user location privacy protection proposed by the present invention is different from the traditional AP deployment method based on positioning error minimization and random selection AP deployment method, the user location in the entire target area is not considered. Cumulative distribution of errors for estimated points vs. results. When considering user location privacy protection, the greater the error of database matching within the reported radius R, the better the system's privacy protection for users. However, for user LBS quality, we expect the error of user location estimation points to be as small as possible while protecting user location privacy.

Detailed ways

The present invention will be further described below in conjunction with accompanying drawing.

The indoor WLAN fingerprint positioning access point deployment method based on user location privacy protection as shown in Figures 1a to 8i includes the following steps:

Step 1: Input the weight coefficient μ, the radius R of the location area reported by the target user to the LBS server, and the unit of R is meter.

Step 2, initialization, let i=1, E_current=0, E_best=0, wherein, i is the count amount, E_current is the objective function value corresponding to the current solution when used to store the optimization search, and E_best is the best value when used to store the optimization search The objective function value corresponding to the optimal solution.

Step 3. According to the number of APs, generate a new AP deployment mode by perturbation, and store the AP candidate location number corresponding to the newly generated AP deployment mode into the matrix slo_new. For the perturbation process, see "Disturbance Generates New AP Deployment Mode" for details. way of way".

Step 4. Use the KNN (K-Nearest Neighbor) algorithm to calculate the position estimation points and corresponding positioning errors of each test point under the current AP deployment mode, and store them in the matrix Location and errors respectively. Among them, the KNN algorithm See "KNN Algorithm Process" for details.

Step 5. All elements in the matrix errors are rounded up and stored in the matrix d

formula one

in, Indicates rounding up.

Step 6, set j=1, where j is the count of the number of test points, and the present invention assumes that the test points are the locations of users who need LBS services.

Step 7. Determine whether the value of d(j,1) is less than R (that is, whether the positioning error of the test point is smaller than the radius of the reported area after rounding up), where d(j,1) is the upward rounding of the positioning error at the test point j The adjusted value; if yes, go to step 8; if not, go to step 15.

Step 8, let k=d(j,1); where k is the counting quantity.

Step 9: Take the estimated point coordinates Location(j,:) of the jth test point as the center of the circle, k as the radius, count the number of users in the circle, calculate the Euclidean distance from different users to the center of the circle, and store it in the matrix Distance .

Step 10. When the attacker obtains the position estimation point of the test point through a certain positioning algorithm (such as the KNN algorithm), considering that there is a certain positioning error, the users in the circle with the position estimation point as the center and the positioning error as the radius are averaged There is a certain probability to select as the target user, assuming that there are U _k users in the circle, then calculate the average information entropy selected as the target user in the circle with a radius of k according to formula 3, and store it in the matrix H (k, 1);

The formula for calculating information entropy is:

formula two

Among them, U _k is the number of people contained in the circle of radius k; p(u) is the probability that the uth person in the circle of radius k is the user; H(u) is the probability of the uth person in the circle of radius k User's information entropy; Distance(u,1) is the Euclidean distance from the uth person in a circle of radius k to the center of the circle.

Since there are U _k individuals in the circle, the average information entropy of users found in the circle with radius k is:

formula three

Step 11. According to Formula 4, calculate the anonymity degree Ad(k,1) of the target user selected by the attacker as the target user in the circle domain with a radius of k when the target user is at the test point j, which reflects the privacy degree of the user:

Ad(k,1)=2 ^H(k,1) Formula 4

Step 12, superimposing Ad(k,1) under different radius k into the matrix element AD(j,1), that is, AD(j,1)=AD(j,1)+Ad(k,1), Among them, AD(j,1) is used to store the total anonymity of the target user selected under different radius k when the target user is at the jth test point.

Step thirteen, make k=k+1, wherein, k equals the radius that the assailant is looking for the target user, because in practical application, about the assailant is unknown for the selection of radius k, therefore according to certain step size (such as etc. Step size 1) Increase the value of k, and calculate the degree of anonymity corresponding to each value of k.

Step 14. Determine whether k is less than or equal to R; if so, go to step 9; otherwise, go to step 16.

Step 15. At this time, the quality of LBS provided by the system will be affected, so define the test point with a positioning error greater than R as the out-of-boundary point, and let the number of out-of-boundary points be r, where r is d(j, 1) (j=1,...,Num_T) is greater than the number of R.

Step sixteen, j=j+1; wherein j is the counted number of test points.

Step 17. Determine whether j is smaller than Num_T; wherein, Num_T is the total number of test points; if yes, go to step 7; if not, go to step 18.

Step eighteen, calculate the average anonymity degree of the user in the current AP deployment mode.

formula five

Among them, Num_T is the total number of test points.

Step nineteen, calculating the average invalidity degree of the users in the current AP deployment mode. Since the out-of-boundary points are test points with positioning errors greater than R, although these test points can obtain better privacy, their LBS service quality is also affected, so we use the average invalidity to describe the impact of LBS service quality, which defines is the average of the number of out-of-bounds points:

formula six

Among them, r is the number of out-of-boundary points; Num_T is the total number of test points.

Step 20. Calculate the objective function value f according to Formula 7:

f＝μ·Aver_AD+(1-μ)·Aver_ID Formula 7

Among them, f is the objective function value under the new AP deployment mode generated by the disturbance; μ is the weight coefficient.

Step 21. Determine whether f is greater than E_current; if so, go to step 22; if not, go to step 25; wherein, E_current is used to store the objective function value corresponding to the current AP deployment mode when optimizing the search;

Step 22, set slo_current=slo_new; E_current=f; wherein slo_current is used to store the current AP deployment mode during the optimization search; slo_new is used to store the new AP deployment mode generated by the optimization search disturbance; E_current is used to store the optimal search time , the objective function value corresponding to the current AP deployment mode.

Step 23: Determine whether f is greater than E_best; if so, go to step 24; if not, go to step 25; wherein, E_best is used to store the objective function value corresponding to the optimal AP deployment mode during optimization search .

Step 24, set slo_best=slo_new; E_best=f; among them, slo_best is used to store the optimal AP deployment method during optimization search; slo_new is used to store the new AP deployment method generated by optimization search disturbance; E_best is used to store optimized When searching, the objective function value corresponding to the optimal AP deployment mode; f is the objective function value under the new AP deployment mode generated by the disturbance.

Step 25, i=i+1; wherein, i is the counting amount.

Step 26. Determine whether i is less than If so, go to step 3; if not, go to step 27; wherein, Num_AP is the total number of APs; AP_candidate is the number of candidate positions for APs; Indicates the number of combinations for selecting Num_AP positions from AP_candidate positions in the permutation combination.

Step 27: Output the optimal AP deployment mode slo_best.

●The method of disturbing the new AP deployment method

The present invention uses an exhaustive search algorithm to search for the optimal AP deployment mode when the number of AP candidate positions is not large. The pseudo code of the new AP deployment mode produced by the disturbance of the present invention is as follows (taking the situation of 3 APs as an example):

1. Set the initial solution

slo_new=[123]; wherein, slo_new is the label used to store the AP candidate position corresponding to the new AP deployment mode;

2. Disturbance generates a new AP deployment method

●KNN algorithm flow

1. Set parameter k;

2. Calculate the distance (such as Euclidean distance) between the user RSS and each reference point RSS in the database;

3. Arrange the distances from small to large;

4. Select the reference points corresponding to the first k distances, and obtain the physical coordinates corresponding to the k reference points;

5. The geometric mean of the physical coordinates of the k reference points is used as the position estimation point coordinates.

As shown in Figure 3, the real environment of the present invention is the aisle on the 5th floor of Shaw Building, Chongqing University of Posts and Telecommunications, which is divided into two areas. There are 38 reference points (marked as " ") in Area 1, and there are 38 reference points (marked as " ") in Area 2. Marked as "·") 15 pieces. In the figure, the five-pointed star symbol represents the origin of the coordinates. The horizontal and vertical coordinate distances of adjacent reference points in area 1 are 2m and 1.2m respectively, and the horizontal and vertical coordinate distances of adjacent reference points in area 2 are 1.2m and 3m respectively. There are 10 The AP candidate positions are labeled 1-10, respectively.

Table 1 shows the physical coordinates of 10 different AP candidate positions and the MAC address of the AP placed at the candidate positions during the test under the real experimental environment of the present invention.

Table 2 shows when the weight coefficient u=0.5 or u=0.7 and the reporting radius R=10m, the optimal AP candidate position, anonymity and invalidity obtained by the method of the present invention, the method based on the minimization of positioning error and the random selection method .

Table 3 shows the optimal AP candidate positions, anonymity and invalidity obtained by the method of the present invention, the method based on positioning error minimization and the random selection method when the weight coefficient u=0.3 and the reporting radius R=10m.

Table 4 shows the optimal AP candidate positions, anonymity and invalidity obtained by the method of the present invention, the method based on positioning error minimization and the random selection method when the weight coefficient u=0.5 and the reporting radius R=5m.

Table I

Table II

Table three

Table four

Claims (5)

1. An indoor WLAN fingerprint positioning access point deployment method based on user position privacy protection is characterized by comprising the following steps:

step one, setting a weight coefficient mu and a position area radius R reported by a target user to an LBS server, wherein the unit of R is meter;

step two, initialization, namely, making i equal to 1, E _ current equal to 0, and E _ best equal to 0, wherein i is a counting number, E _ current is used for storing an objective function value corresponding to a current solution during optimization search, and E _ best is used for storing an objective function value corresponding to an optimal solution during optimization search;

step three, generating a new AP deployment mode in an interference mode according to the number of the APs, and storing the AP candidate position mark number corresponding to the newly generated AP deployment mode into a matrix slo _ new;

calculating position estimation points and corresponding positioning errors of all test points in the current AP deployment mode by using a KNN algorithm, and respectively storing the position estimation points and the corresponding positioning errors into a matrix Location and errors;

step five, rounding all elements in the matrix error upwards and storing the rounded elements in the matrix d;

step six, making j equal to 1, wherein j is the counting quantity of the number of the test points, and supposing that the test points are the position points of the users needing LBS;

step seven, judging whether the value of d (j,1) is smaller than R, namely whether the positioning error of the test point is smaller than the reported position area radius R after being rounded upwards, wherein d (j,1) is the value of the positioning error of the test point j after being rounded upwards; if yes, entering step eight; if not, entering the step fifteen;

step eight, making k equal to d (j, 1); wherein k is a count;

step nine, taking the position estimation point coordinate Location (j,: of the jth test point) as the circle center and k as the radius, counting the number of users in the circle domain, calculating the Euclidean distances from different users to the circle center, and storing the Euclidean distances into a matrix Distance;

step ten, supposing that U exists in the circular domain_kCalculating the average information entropy of the target user selected in the circular domain with the radius of k and storing the average information entropy into a matrix H (k, 1);

step eleven, when the target user is at the test point j, an attacker selects the anonymity Ad (k,1) of the target user in a circular domain with the radius of k, and the value reflects the privacy degree of the user;

step twelve, adding ADs (k,1) with different radii k to the matrix element Ad (j,1), that is, Ad (j,1) ═ Ad (j,1) + Ad (k,1), where Ad (j,1) is used to store the total anonymity of the target user selected under different radii k at the jth test point;

step thirteen, making k equal to k +1, wherein k is equal to the radius of the target user searched by the attacker, and calculating the corresponding anonymity degree of each k value;

step fourteen, judging whether k is less than or equal to R, if so, entering step nine; if not, entering the step sixteen;

fifteenth, defining the test points with the positioning errors larger than R as boundary points, and enabling the number of the boundary points to be R, wherein R is d (j,1) (j is 1, …, Num _ T) and the value of R is larger than the number of R;

sixthly, making j equal to j + 1;

seventhly, judging whether j is smaller than Num _ T, wherein Num _ T is the total number of the test points; if yes, entering a seventh step; if not, entering the eighteen steps;

eighteen, calculating the average anonymity of the user in the current AP deployment mode;

nineteenth, calculating the average inefficiency of the user in the current AP deployment mode, wherein the calculation formula of the average inefficiency is as follows:

<mrow> <mi>A</mi> <mi>v</mi> <mi>e</mi> <mi>r</mi> <mo>_</mo> <mi>I</mi> <mi>D</mi> <mo>=</mo> <mfrac> <mi>r</mi> <mrow> <mi>N</mi> <mi>u</mi> <mi>m</mi> <mo>_</mo> <mi>T</mi> </mrow> </mfrac> </mrow>

wherein r is the number of the out-of-bounds points; num _ T is the total number of test points;

twenty, calculating an objective function value f, wherein f is the objective function value generated by disturbance in a new AP deployment mode; the calculation formula of the objective function value f is as follows:

f＝μ·Aver_AD+(1-μ)·Aver_ID

wherein, f is an objective function value generated by disturbance in a new AP deployment mode; mu is a weight coefficient;

twenty one, judging whether f is larger than E _ current; if yes, entering the twenty-two step; if not, entering twenty-five step;

twenty-two, let slo _ current be slo _ new; f, wherein slo _ current is used for storing a current AP deployment mode during optimized search; slo _ new is used for storing a new AP deployment mode obtained due to optimization search disturbance;

twenty-third, judging whether f is larger than E _ best; if yes, entering twenty-four steps; if not, entering twenty-five step;

twenty-four, let slo _ best be slo _ new, E _ best be f; the slo _ best is used for storing an optimal AP deployment mode during optimized search;

twenty five, making i equal to i + 1;

twenty-six step, judging whether i is less thanIf yes, entering a third step; if not, entering a twenty-seventh step, wherein Num _ AP is the total number of the APs, and AP _ candidate is the number of the AP candidate positions;a number of combinations representing Num _ AP different AP positions selected from AP _ candidate AP positions in a permutation and combination manner;

and twenty-seven, outputting slo _ best.

2. The method for deploying the indoor WLAN fingerprint positioning access point based on the user location privacy protection as claimed in claim 1, wherein: in the fifth step, the matrix d is:

wherein,indicating rounding up.

3. The method for deploying the indoor WLAN fingerprint positioning access point based on the user location privacy protection as claimed in claim 2, wherein: in the step ten, the calculation formula of the average information entropy is:

<mrow> <mi>H</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>,</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>u</mi> <mo>=</mo> <mn>1</mn> </mrow> <msub> <mi>U</mi> <mi>k</mi> </msub> </munderover> <mi>H</mi> <mrow> <mo>(</mo> <mi>u</mi> <mo>)</mo> </mrow> </mrow> <msub> <mi>U</mi> <mi>k</mi> </msub> </mfrac> </mrow>

the information entropy H (u) is calculated by the formula:

<mfenced open = "{" close = ""> <mtable> <mtr> <mtd> <mi>H</mi> <mo>(</mo> <mi>u</mi> <mo>)</mo> <mo>=</mo> <mo>-</mo> <mi>p</mi> <mo>(</mo> <mi>u</mi> <mo>)</mo> <mo>&amp;times;</mo> <mi>l</mi> <mi>o</mi> <msub> <mi>g</mi> <mn>2</mn> </msub> <mi>p</mi> <mo>(</mo> <mi>u</mi> <mo>)</mo> </mtd> </mtr> <mtr> <mtd> <mi>p</mi> <mo>(</mo> <mi>u</mi> <mo>)</mo> <mo>=</mo> <mfrac> <mrow> <mi>D</mi> <mi>i</mi> <mi>s</mi> <mi>tan</mi> <mi>c</mi> <mi>e</mi> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>u</mi> <mo>=</mo> <mn>1</mn> </mrow> <msub> <mi>U</mi> <mi>k</mi> </msub> </munderover> <mi>D</mi> <mi>i</mi> <mi>s</mi> <mi>tan</mi> <mi>c</mi> <mi>e</mi> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> </mfrac> </mtd> </mtr> </mtable> </mfenced>

wherein, U_kThe number of persons contained in a circle of radius k; p (u) is the probability that the u-th person within the circle of radius k is a user; h (u) is the information entropy that the u th person in the circle with the radius k is the user; distance (u,1) is the Euclidean Distance from the center of the circle of radius k for the u-th person.

4. The method for deploying the indoor WLAN fingerprint positioning access point based on the user location privacy protection as claimed in claim 3, wherein: in the eleventh step, a calculation formula of the anonymity degree Ad (k,1) of the target user is as follows:

Ad(k,1)＝2^H(k,1)。

5. the method for deploying the indoor WLAN fingerprint positioning access point based on the user location privacy protection as claimed in claim 4, wherein: in the eighteenth step, the formula for calculating the average anonymity of the user in the current AP deployment mode is as follows:

<mrow> <mi>A</mi> <mi>v</mi> <mi>e</mi> <mi>r</mi> <mo>_</mo> <mi>A</mi> <mi>D</mi> <mo>=</mo> <mfrac> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mrow> <mi>N</mi> <mi>u</mi> <mi>m</mi> <mo>_</mo> <mi>T</mi> </mrow> </munderover> <mi>A</mi> <mi>D</mi> <mrow> <mo>(</mo> <mi>j</mi> <mo>,</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> <mrow> <mi>N</mi> <mi>m</mi> <mi>u</mi> <mo>_</mo> <mi>T</mi> </mrow> </mfrac> </mrow>

wherein Num _ T is the total number of test points.