CN109587627A - The indoor positioning algorithms of terminal heterogeneity are improved based on RSSI - Google Patents

Abstract

The invention discloses an indoor positioning algorithm for improving terminal heterogeneity problem based on RSSI, comprising the following steps. Step (1): Determine each test point in the test space, and calculate the distance from each test point to each transmitting terminal. Step (2): Make the same combination of multiple distance data obtained from each test point, and the data in the number of combinations must be greater than or equal to 3. The indoor positioning algorithm based on RSSI to improve the problem of terminal heterogeneity disclosed by the invention does not need to know the values of A and n when the data is standardized, and theoretically eliminates the influence of equipment differences and environmental differences. ;Fingerprint data is obtained by processing the distance from the fingerprint coordinate point to each sender, and does not require manual collection, which reduces labor loss; standardized processing will reduce the speed of positioning, and fingerprint data is prepared in advance, to a certain extent The real-time positioning is guaranteed.

Description

Improved indoor positioning algorithm for terminal heterogeneity problem based on RSSI

technical field

The invention belongs to the technical field of wireless signal positioning, in particular to an indoor positioning algorithm based on RSSI to improve the problem of terminal heterogeneity.

Background technique

Traditional RSSI-based indoor positioning algorithms, such as fingerprint algorithm, maximum likelihood estimation algorithm, trilateral positioning algorithm, minimum maximum value algorithm, etc. The RSSI ranging principle is used in these algorithms. The relationship between the received signal strength of the wireless signal and the signal transmission distance is expressed by formula (1), where RSSI is the received signal strength, and d is the distance between the transmitter and the receiver. , n is the signal propagation factor, A is the absolute value of the signal strength received by the receiver when the transmitter and receiver are separated by 1 meter.

It can be seen from the formula (1) that the values of the constants A and n determine the relationship between the received signal strength and the signal transmission distance.

Among them, the maximum likelihood estimation algorithm, the trilateral positioning algorithm and the minimum maximum value algorithm are all based on the formula (1). When the signal transmission distance is determined, the corresponding algorithm is used for positioning, that is, in different environments And in different terminal situations, it is necessary to continuously debug the values of A and n to better locate. The fingerprint algorithm is different from the other three algorithms, but uses the matching between RSSIs. The method has the following steps:

1. Through a receiving terminal, each test point in the test space receives the RSSI of each transmitter in the space as fingerprint data;

2. Receive the RSSI of each transmitter as the test data at a certain position in the test space through a receiver terminal that needs to be positioned;

3. The test data and the fingerprint data are matched by the k-nearest neighbor (knn) algorithm, and the test point with the best matching is the positioning position.

The algorithm does not completely depend on formula (1), is not affected by the complex environment, and has high accuracy, but the algorithm is very labor-intensive compared with other algorithms, and the fingerprint data is more dependent on a certain receiver. Depending on the model of the terminal, the corresponding fingerprint data will be different; or if the environment when the fingerprint is collected is changed, the corresponding fingerprint data will also be different. The invention improves the data processing on the basis of the fingerprint algorithm, and is not affected by the environment and different receiving terminals under the premise of not needing labor.

SUMMARY OF THE INVENTION

Aiming at the shortcomings of the background technology, the present invention overcomes the above shortcomings, and provides an indoor positioning algorithm based on RSSI to improve the problem of terminal heterogeneity. to improve the positioning accuracy.

The present invention adopts the following technical solutions, and the indoor positioning algorithm for improving terminal heterogeneity problem based on RSSI includes the following steps:

Step (1): determine each test point in the test space, calculate the distance from each test point to each transmitting terminal;

Step (2): Make the same combination of multiple distance data obtained from each test point, and the data in the number of combinations must be greater than or equal to 3;

Step (3): each test point selects the same combination, and under this combination, the distance data of each test point is first processed logarithmically, and then standardized, and the obtained data is used as the fingerprint data under this combination, wherein each The processed data of each test point under a combination is used as fingerprint data;

Step (4): collect the signal strength values of three or more transmitting ends through the receiving end at the position point that needs to be located, and standardize the absolute value of the collected signal strength values as test data;

Step (5): select the corresponding combination of fingerprint data according to the serial number of the transmitter corresponding to the test data, and use the k nearest neighbor algorithm to match the test data and the fingerprint data, and the matched optimal position point is the desired positioning point.

According to the above technical solution, the distance described in step (1) is d_ij, where i is the ith position, j is the jth transmitter, and d_ij represents the distance from the ith position to the jth transmitter.

According to the above technical solution, the specific situation of combining the data in step (2) is: assuming that there are 4 receiving terminals in a certain test space, then each position point to each receiving terminal will have 4 distances d _i1 , d _i2 , d _i3 , d _i4 , where i is the i-th position point, 1, 2, 3, and 4 are the serial numbers of the transmitters, and the following 5 combinations are obtained: (d _i1 , d _i2 , d _i3 ), (d _i1 , d _i2 , d _i4 ), (d _i1 , d _i3 , d _i4 ), (d _i2 , d _i3 , d _i4 ), (d _i1 , d _i2 , d _i3 , d _i4 ).

According to above-mentioned technical scheme, in step (3), the principle steps of standardizing data are as follows:

Step a. It can be known from the distance formula of signal strength: And the value of A, n is fixed but unknown, then the right part of the equation is converted into a linear function:

in, are constant values;

Step b. Understand the properties of mathematical expectation and variance, as well as standardized algorithm formulas;

The property of mathematical expectation: if there is a random variable X, for any constant a, b, there are:

E(aX+b)=aE(X)+b,

The nature of variance: if there is a random variable X, for any constant a, b, there are:

Var(aX+b)=a ² Var(X),

The corresponding standard deviation is:

Standardized algorithm formula:

Among them, _xi is the original data, _zi is the new standardized data, is the mean of the original data, and s is the standard deviation of the original data;

Step c. Select the RSSI of 4 transmitters obtained at a certain position as a set of data (rssi ₁ , rssi ₂ , rssi ₃ , rssi ₄ ), and link the linear function of step a with the properties of step b and the standardized formula to obtain:

Among them, m and s are the mean and standard deviation of the group of RSSI data, i=1, 2, 3, 4;

Step d. Corresponding to the RSSI selected in step c, obtain:

According to the above-mentioned technical scheme, the fingerprint data source steps in step (3) are as follows:

The first step, (for the situation in claim 3) select a combination (d _i1 , d _i2 , d _i3 ) from 5 combinations;

The second step is to obtain (log ₁₀ d _i1 , log ₁₀ d _i2 , log ₁₀ d _i3 ) after logarithmic processing;

The third step is to standardize the logarithmically processed data to obtain:

Among them, dm and ds are the mean and standard deviation of (log ₁₀ d _i1 , log ₁₀ d _i2 , log ₁₀ d _i3 ), respectively. As the fingerprint data of the i-th position point position under the combination;

The fourth step is to repeat the second step and the third step for each position point in the test space to obtain the fingerprint data of each position point as the fingerprint data of the entire test space under the combination;

In the fifth step, the fourth step is repeated for each combination, and the fingerprint data under each combination is obtained as the final fingerprint data.

According to above-mentioned technical scheme, the data matching step in step (5) is as follows:

(i). Obtain all fingerprint data according to claim 5;

(ii). Receive the RSSI of each sender at the position that needs to be positioned, and normalize the absolute value of this group of RSSI to obtain a new data group as the test data;

(iii). Find the fingerprint data in the corresponding combination of all fingerprint data according to the RSSI sender serial number received in (ii) as offline data

(iv). Calculate the distance between the new data group and the test data under each coordinate point in the offline data;

(ⅴ). Sort in ascending order of distance, and select k coordinate points with the smallest distance;

(ⅵ). Under this method, k=1 is selected, and the coordinate point with the smallest distance between offline data and test data is used as the coordinate point obtained by positioning.

The indoor positioning algorithm based on RSSI to improve the terminal heterogeneity problem disclosed by the present invention has the beneficial effect that, firstly, it is no longer necessary to know the values of A and n when the data is standardized, which theoretically eliminates equipment differences and The impact of environmental differences; secondly, the fingerprint data is obtained by processing the distance from the fingerprint coordinate point to each sender, which does not require manual collection, which reduces manual loss; in addition, the standardized processing will reduce the speed of positioning, The fingerprint data is prepared in advance to ensure the real-time positioning to a certain extent.

Description of drawings

Figure 1 is a basic logic flow diagram of the present invention.

FIG. 2 is a schematic diagram of offline fingerprint data of the present invention.

FIG. 3 is a schematic diagram of the average positioning error of different cases A under each algorithm of the present invention.

FIG. 4 is a schematic diagram of the average positioning error of different situations n under each algorithm of the present invention.

Figure 5 is a schematic diagram of the main steps of the present invention.

Detailed ways

The present invention discloses an indoor positioning algorithm based on RSSI to improve the terminal heterogeneity problem. The specific implementation of the present invention will be further described below with reference to the preferred embodiments.

Referring to FIG. 1 to FIG. 5 of the accompanying drawings, FIG. 1 shows the basic processing logic of the indoor positioning algorithm based on the RSSI-based improvement of the terminal heterogeneity problem, and FIG. 2 shows the RSSI-based improvement of the terminal heterogeneity problem. The offline fingerprint data of the indoor positioning algorithm, Figure 3 shows the average positioning error of each algorithm of the indoor positioning algorithm based on the RSSI improvement of the terminal heterogeneity problem under different conditions, and Figure 4 shows the improvement based on RSSI. The average positioning error of each algorithm of the indoor positioning algorithm for the terminal heterogeneity problem under different conditions, Figure 5 shows the main steps of the indoor positioning algorithm based on the RSSI improvement for the terminal heterogeneity problem.

Preferably, according to the above preferred embodiment, the RSSI-based indoor positioning algorithm for improving terminal heterogeneity includes the following steps:

Step (1): determine each test point in the test space, calculate the distance from each test point to each transmitting terminal;

Step (2): Make the same combination of multiple distance data obtained from each test point, and the data in the number of combinations must be greater than or equal to 3;

Step (3): each test point selects the same combination, and under this combination, the distance data of each test point is first processed logarithmically, and then standardized, and the obtained data is used as the fingerprint data under this combination, wherein each The processed data of each test point under a combination is used as fingerprint data;

Step (4): collect the signal strength values of three or more transmitting ends through the receiving end at the position point that needs to be located, and standardize the absolute value of the collected signal strength values as test data;

Step (5): compare the serial number of the transmitter corresponding to the test data, select the fingerprint data of the corresponding combination, and use the k nearest neighbor (knn) algorithm to match the test data and the fingerprint data, and the matched optimal position point is the desired location. point.

Further, the distance described in step (1) is d _ij , where i is the ith position point, j is the j th transmitting end, and d _ij represents the distance from the ith position point to the j th transmitting end.

Further, the specific situation of combining the data in step (2) is: assuming that there are 4 receiving terminals in a certain test space, then each position point to each receiving terminal will have 4 distances d _i1 , d _i2 respectively, d _i3 , d _i4 , where i is the ith position point, 1, 2, 3, 4 are the serial numbers of the transmitter, and the following 5 combinations are obtained: (d _i1 , d _i2 , d _i3 ), (d _i1 , d _i2 , d _i4 ), (d _i1 , d _i3 , d _i4 ), (d _i2 , d _i3 , d _i4 ), (d _i1 , d _i2 , d _i3 , d _i4 ).

Further, in step (3), the principle steps of standardizing the data are as follows:

Step a. It can be known from the distance formula of signal strength: And the value of A, n is fixed but unknown, then the right part of the equation is converted into a linear function:

in, are constant values;

Step b. Understand the properties of mathematical expectation and variance, as well as standardized algorithm formulas;

The property of mathematical expectation: if there is a random variable X, for any constant a, b, there are:

E(aX+b)=aE(X)+b,

The nature of variance: if there is a random variable X, for any constant a, b, there are:

Var(aX+b)=a ² Var(X),

The corresponding standard deviation is:

Standardized algorithm formula:

Among them, _xi is the original data, _zi is the new standardized data, is the mean of the original data, and s is the standard deviation of the original data;

Step c. Select the RSSI of 4 transmitters obtained at a certain position as a set of data (rssi ₁ , rssi ₂ , rssi ₃ , rssi ₄ ), and link the linear function of step a with the properties of step b and the standardized formula to obtain:

Among them, m and s are the mean and standard deviation of the group of RSSI data, i=1, 2, 3, 4;

Step d. Corresponding to the RSSI selected in step c, obtain:

At the same time, it can also be seen from step c that for the data set The new data obtained by normalizing and normalizing the data set |rssi _i | are equal, and then according to the above four equations. That is, normalizing the data sets (log ₁₀ d ₁ , log ₁₀ d ₂ , log ₁₀ d ₃ , log ₁₀ d ₄ ) and normalizing the data sets (|rssi ₁ |, |rssi ₂ |, |rssi ₃ |, |rssi ₄ |) and the new data obtained by normalization are the same.

Further, the fingerprint data source step in step (3) is as follows:

The first step, (for the situation in claim 3) select a combination (d _i1 , d _i2 , d _i3 ) from 5 combinations;

The second step is to obtain (log ₁₀ d _i1 , log ₁₀ d _i2 , log ₁₀ d _i3 ) after logarithmic processing;

The third step is to standardize the logarithmically processed data to obtain:

Among them, dm and ds are the mean and standard deviation of (log ₁₀ d _i1 , log ₁₀ d _i2 , log ₁₀ d _i3 ), respectively. As the fingerprint data of the i-th position point position under the combination;

The fourth step is to repeat the second step and the third step for each position point in the test space to obtain the fingerprint data of each position point as the fingerprint data of the entire test space under the combination;

In the fifth step, the fourth step is repeated for each combination, and the fingerprint data under each combination is obtained as the final fingerprint data.

Further, the data matching step in step (5) is as follows:

(i). Obtain all fingerprint data according to claim 5;

(ii). Receive the RSSI of each sender at the position that needs to be positioned, and normalize the absolute value of this group of RSSI to obtain a new data group as the test data;

(iii). Find the fingerprint data in the corresponding combination of all fingerprint data according to the RSSI sender serial number received in (ii) as offline data

(iv). Calculate the distance between the new data group and the test data under each coordinate point in the offline data;

(ⅴ). Sort in ascending order of distance, and select k coordinate points with the smallest distance;

(ⅵ). Under this method, k=1 is selected, and the coordinate point with the smallest distance between offline data and test data is used as the coordinate point obtained by positioning.

It is worth mentioning that, according to a variant of the above preferred embodiment, the indoor positioning algorithm based on RSSI to improve the terminal heterogeneity problem may further include the following steps:

Step 1: Interact with the AP, obtain the AP and RSSI data corresponding to the terminal, and perform preprocessing operations including but not limited to filtering RSSI and grouping;

Step 2: According to the obtained AP information, select the coordinates of the area where the AP is located, and use the method mentioned in this article to construct offline fingerprint data. The specific construction method is as follows:

(2.1) Obtain the test space area where the fingerprint needs to be constructed according to the ap coordinates, and grid the test space area;

(2.2) For each grid, adopt the formula mentioned above, and use the distance from the center point of the grid to the AP to carry out the standardization process to obtain offline fingerprint data;

Step 3: standardize the absolute value of RSSI data corresponding to ap and bring it into the fingerprint information, perform KNN calculation, and obtain the coordinate point with the smallest data distance;

Step 4: Perform secondary processing on the obtained coordinates, including but not limited to path optimization, point filtering and other operations, and output final coordinate values.

According to a variant embodiment of the above-mentioned preferred embodiment, one of the above-mentioned algorithms is described as follows.

It should be noted that in the above step description, for the convenience of description, the offline fingerprint construction operation is performed after the signal strength data is obtained, and because the construction of the offline fingerprint data itself will consume more resources, in the actual production environment Most of the possible combinations of offline fingerprint data are often pre-built in the fingerprint system, and a small part of the fingerprint data that cannot be cached will be cached after being generated once. Thereby greatly improving the running speed of the algorithm. The final running speed is comparable to that of traditional localization algorithms.

At the same time, since the average error of the maximum likelihood estimation method and the trilateral positioning algorithm is very large under different A and n cases, the two algorithms are not displayed and compared in Figure 3 and Figure 4. From Figure 3 and Figure 4, it can be seen that the indoor positioning algorithm based on RSSI to improve the terminal heterogeneity problem has little change in the overall average positioning error under different A and n conditions, effectively reducing the problems caused by different environments and different equipment. error to come.

For those skilled in the art, the technical solutions described in the foregoing embodiments can still be modified, or some technical features thereof can be equivalently replaced. Any modifications made within the spirit and principles of the present invention, Equivalent replacements, improvements, etc., should all be included in the protection scope of the present invention.

Claims (6)

1. An indoor positioning algorithm for improving the problem of terminal heterogeneity based on RSSI (received signal strength indicator) is characterized by comprising the following steps:

step (1): determining each test point in the test space, and calculating the distance from each test point to each sending terminal;

step (2): the same combination is carried out on a plurality of distance data obtained by each test point, and the data in the combination number is required to be more than or equal to 3;

and (3): selecting the same combination from each test point, and under the combination, firstly carrying out logarithmic processing on the distance data of each test point, then carrying out standardized processing on the distance data to obtain data serving as fingerprint data under the combination, wherein the processed data of each test point under each combination is used as the fingerprint data;

and (4): acquiring signal intensity values of three or more transmitting ends at a position point to be positioned through a receiving end, and standardizing absolute values of the acquired signal intensity values to be used as test data;

and (5): and comparing the sending end serial numbers corresponding to the test data, selecting the fingerprint data of the corresponding combination, and matching the test data and the fingerprint data by using a k-nearest neighbor algorithm, wherein the matched optimal position point is the solved positioning point.

2. The indoor positioning algorithm for improving terminal heterogeneity problem based on RSSI of claim 1, wherein the distance in step (1) is d _ ij, where i is the ith position point, j is the jth sender, and d _ ij represents the distance from the ith position point to the jth sender.

3. The indoor positioning algorithm for improving terminal heterogeneity problem based on RSSI as claimed in claim 2, wherein the step (2) combines the data as follows: assuming that there are 4 receiving terminals in a certain test space, there are 4 distances d from each position point to each receiving terminal respectively_i1，d_i2，d_i3，d_i4Wherein i is the ith position point, 1,2,3,4 are the serial number of the transmitting terminal, and the following 5 combinations are obtained: (d)_i1，d_i2，d_i3)，(d_i1，d_i2，d_i4)，(d_i1，d_i3，d_i4)，(d_i2，d_i3，d_i4)，(d_i1，d_i2，d_i3，d_i4)。

4. The indoor positioning algorithm for improving terminal heterogeneity problem based on RSSI of claim 3, wherein the principle steps of normalizing the data in step (3) are as follows:

step a, obtaining a distance formula of signal intensity:and the value of a, n is fixed but unknown, then the right part of the equation is converted to a linear function:

wherein,are all constant values;

b, knowing the nature of mathematical expectation and variance, and a standardized algorithm formula;

mathematically desirable properties: if there is a random variable X, for any constant a, b, there are:

E(aX+b)＝aE(X)+b，

the nature of the variance: if there is a random variable X, for any constant a, b, there are:

Var(aX+b)＝a²Var(X)，

the corresponding standard deviations are:

a normalized algorithm formula:

wherein x is_iAs raw data, z_iIn order to be able to normalize the new data,the average value of the original data is obtained, and s is the standard deviation of the original data;

c, selecting RSSI of 4 transmitting terminals obtained at a certain position as a group of numbersAccording to (rssi)₁，rssi₂，rssi₃，rssi₄) And connecting the linear function of the step a and the property of the step b and a standardized formula to obtain:

wherein m and s are the mean and standard deviation of the set of RSSI data, respectively, and i is 1,2,3, 4;

and d, corresponding to the RSSI selected in the step c, obtaining:

5. the indoor positioning algorithm for improving terminal heterogeneity problem based on RSSI of claim 3, wherein the fingerprint data source step in step (3) is as follows:

in a first step, a combination (d) is selected from 5 combinations (for the case of claim 3)_i1，d_i2，d_i3)；

Second, obtained after logarithmic treatment (log)₁₀d_i1，log₁₀d_i2，log₁₀d_i3)；

Thirdly, carrying out standardization processing on the data after logarithmic processing to obtain:

wherein dm and ds are (log)₁₀d_i1，log₁₀d_i2，log₁₀d_i3) Average and standard deviation ofAs fingerprint data of the ith position point under the combination;

fourthly, repeating the second step and the third step for each position point in the test space to obtain the fingerprint data of each position point, and taking the fingerprint data as the fingerprint data of the whole test space under the combination;

and fifthly, repeating the fourth step for each combination to obtain the fingerprint data under each combination as final fingerprint data.

6. The indoor positioning algorithm for improving terminal heterogeneity problem based on RSSI of claim 3, wherein the data matching step in step (5) is as follows:

obtaining all fingerprint data according to claim 5;

(ii) receiving the RSSI of each sending end at a position point needing positioning, and carrying out standardization processing on the absolute values of the RSSI to obtain a new data set as test data;

finding fingerprint data under corresponding combination in all fingerprint data as offline data aiming at sending end serial numbers of received RSSI (received signal strength indicator) in (ii)

(iv) calculating the distance between the new data group and the test data under each coordinate point in the offline data;

(v) sorting according to the distance increasing order, and selecting k coordinate points with the minimum distance;

and (vi) selecting k as 1 under the method, and taking the coordinate point with the minimum distance between the offline data and the test data as the coordinate point obtained by positioning.