CN109444812B - A RSSI Indoor Localization Method Introducing Dynamic Threshold - Google Patents

Abstract

The invention discloses an RSSI indoor positioning method which introduces a dynamic threshold. A one-dimensional linear regression equation is established to calculate the signal strength value of the newly added virtual reference label in the border area, and the virtual reference label with the same density distribution as the central area is added; The boundary line is a mirror surface, and the reference label in the positioning area is mapped to the other side of the boundary line to become the mirror reference label and the mirror virtual label. Find the RSSI value of the reference label with the largest RSSI value difference from the label to be tested to obtain the difference between the two. Absolute value, modify the unified threshold used when constructing the proximity map to a dynamic threshold that changes according to the different labels to be tested, and then construct a proximity map based on the dynamic threshold to filter out the reference labels suitable for the labels to be tested, and according to the weight value of the reference label and its value, The coordinates solve the position of the label to be measured, and complete the positioning of the target in the entire area to be measured. The present invention improves the positioning accuracy at the boundary through the newly added virtual reference label.

Description

A RSSI Indoor Localization Method Introducing Dynamic Threshold

technical field

The invention belongs to the technical field of radio frequency identification (RFID) communication, and in particular relates to an RSSI indoor positioning method that introduces a dynamic threshold.

Background technique

Although the mature GPS (Global Positioning System) is widely used in the outdoor environment, its positioning effect in the indoor environment is not satisfactory, because the wireless signal transmitted by the satellite cannot be effectively utilized in the indoor environment. Radio Frequency Identification (RFID) technology has entered the field of indoor positioning with many advantages such as non-contact, non-line-of-sight, low cost, large storage capacity, fast recognition speed, strong anti-interference ability and good security.

Among various RFID-based indoor positioning methods, it is found that the indoor positioning system using Received Signal Strength Indication (RSSI) has the characteristics of low cost and easy implementation, and it can achieve a certain degree of positioning accuracy. It has gradually become a more practical indoor positioning method at this stage. VIRE (Active RFID-based Localization Using Virtual Reference Elimination) is an indoor positioning algorithm based on RSSI. It linearly inserts virtual reference tags in the positioning area and constructs a proximity map to exclude the small probability positions of the tags to be tested. Improved the positioning effect of the LANDMARC (Indoor Location Sensing Using Active RFID) system. Nonetheless, VIRE still has its shortcomings. Since the newly added virtual reference label only covers the central part of the positioning area, the positioning effect of the point to be measured at the boundary is not as good as that at the center; in addition, the overall positioning accuracy of VIRE still has room for further improvement. Therefore, the present invention focuses on the problem of how to improve the positioning accuracy of the VIRE algorithm.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to provide an RSSI indoor positioning method that introduces dynamic thresholds in view of the above-mentioned deficiencies in the prior art. The four boundary lines of the positioning area are used as mirror surfaces. Various reference labels in the half range are symmetrically mapped outside the boundary of the positioning area, and the positioning accuracy at the boundary is improved by these newly added virtual reference labels.

The present invention adopts following technical scheme:

An RSSI indoor positioning method that introduces dynamic thresholds. The VIRE algorithm is used to establish a one-dimensional linear regression equation of the signal strength values and coordinate values of the actual reference tags and virtual reference tags, and the signal strength values of the newly added virtual reference tags in the border area are calculated. The positioning boundary area is added with virtual reference labels with the same density distribution as the central area; then the positioning boundary line is used as a mirror surface, and the reference label in the positioning area is mapped to the other side of the boundary line to become a mirror reference label and a mirror virtual label, so that the boundary area is indirectly It becomes a central area surrounded by virtual tags; according to the RSSI value of the tag to be tested and the RSSI value of the reference tag read by the reader, find out the RSSI value of the reference tag with the largest difference with the RSSI value of the tag to be tested to obtain the difference between the two. Absolute value, determine the threshold used when constructing a proximity map for the label to be tested, and modify the unified threshold used when building a proximity map to a dynamic threshold that changes according to the label to be tested; For the reference label of the label to be measured, the position of the label to be measured is obtained according to the weight value of the reference label and its coordinates, so as to complete the positioning of the target in the entire area to be measured.

Specifically, it includes the following steps:

S1. Four readers are arranged at the four corners of the positioning area, and virtual reference tags of N*N structure are evenly distributed in the square formed by every four actual reference tags as the top corners. According to the actual reference tags read by the readers The RSSI of the N*N distributed virtual reference label calculated by the linear interpolation method mentioned in VIRE establishes a one-dimensional linear regression equation;

S2, calculate the signal strength value of each boundary virtual reference label newly added on the boundary according to the one-dimensional linear regression equation established in step S1;

S3. Using the four boundary lines as mirror surfaces, the labels in nearly half of the area to be measured inside the boundary line are symmetrically mapped to the outside of the boundary line, and the actual reference label and the virtual reference label in the area to be measured are respectively mapped to the mirror image reference in the mirror area Tag and mirror virtual tag, calculate the RSSI value of the mirror tag read by the reader in the system;

S4. According to the RSSI value of the tag to be tested and the RSSI value of the reference tag read by the reader, find out the RSSI value of the reference tag with the largest difference with the RSSI value of the tag to be tested, and calculate the absolute value of the difference between the two to determine The threshold used when the tag to be tested constructs the proximity map;

S5, constructing a proximity map according to the threshold to filter out the reference labels suitable for the label to be tested, and solve the weights of these reference labels, and then use the weight value and its coordinates to solve the estimated position (x', y') of the label to be measured;

S6. If the first estimated position (x', y') of the tag to be tested is located in the boundary area, multiply the absolute value in step S4 by 0.35 as the secondary threshold used when the tag to be tested constructs the proximity map, and then use the new threshold Carry out the steps of constructing a neighborhood map, finding weight factors and estimating the position of the label to be measured; if the first estimated position (x', y') of the label to be measured is located in the central area, the original threshold and the positioning result are not changed.

Further, in step S1, the one-dimensional linear equation is as follows:

x为标签所处位置的横坐标，s为阅读器读取到标签的RSSI值，设系统中阅读器个数为n，四个实际参考标签和多个虚拟参考标签在阅读器i上的RSSI值分别为S_i1，S_i2，S_i3，S_i4和s_i1，s_i2，s_i3，s_i4，s_i5，s_i6，1≤i≤n，在系统中所处位置的横坐标分别为X₁，X₂，X₃，X₄和x₁，x₂，x₃，x₄，x₅，x₆，RSSI值和横坐标数据组为(X₁,S_i1)，(X₂,S_i2)，(X₃,S_i3)，(X₄,S_i4)和(x₁,s_i1)，(x₂,s_i2)，(x₃,s_i3)，(x₄,s_i4)，(x₅,s_i5)，(x₆,s_i6)。in,

x is the abscissa of the location of the tag, s is the RSSI value of the tag read by the reader, and the number of readers in the system is n, the RSSI of four actual reference tags and multiple virtual reference tags on reader i The values are S _i1 , S _i2 , S _i3 , S _i4 and s _i1 , s _i2 , s _i3 , s _i4 , s _i5 , s _i6 , 1≤i≤n, and the abscissas of the positions in the system are respectively X ₁ , X ₂ , X ₃ , X ₄ and x ₁ , x ₂ , x ₃ , x ₄ , x ₅ , x ₆ , the RSSI value and the abscissa data set are (X ₁ , S _i1 ), (X ₂ , S _i2 ), (X ₃ ,S _i3 ),(X ₄ ,S _i4 ) and (x ₁ ,s _i1 ),(x ₂ ,s _i2 ),(x ₃ ,s _i3 ),(x ₄ ,s _i4 ) ), (x ₅ ,s _i5 ), (x ₆ ,s _i6 ).

Further, in step S2, a linear interpolation method is used to obtain the RSSI value of the newly added virtual reference tag in the part between the four boundary lines and the outermost periphery of the central area.

Further, the linear interpolation formula in the horizontal direction is as follows:

The linear interpolation formula in the vertical direction is as follows:

0≤p＝i％n≤n-1，0≤q＝j％n≤n-1。Among them, _Sk (T _i,j ) represents the RSSI value of the virtual reference tag at (i,j) read by the kth reader,

0≤p=i%n≤n-1, 0≤q=j%n≤n-1.

Further, in step S3, the upper, lower, left, and right boundary lines are respectively used as mirror surfaces, and the mirror label outside the boundary is added to the system through the mapping function, as follows:

When the x-axis is used as the mirror surface, the actual reference labels in the positioning area are counted in the order from left to right and bottom to top, until it reaches half of the total number of actual reference labels in the system. The straight line represented by the ordinate is used as the upper boundary of the area to be mirrored; the x-axis, y-axis and the right boundary of the positioning area are respectively used as the lower boundary, left boundary and right boundary of the area to be mirrored; the actual reference label in the area to be mirrored And the virtual reference label is mapped to the corresponding mirror image area below the x-axis according to the mirror imaging principle to become the mirror reference label and the mirror virtual label; the position coordinates of the label in the mirror area and the label in the to-be-mirrored area before the mapping are symmetrical about the x-axis, respectively for images and things;

When the y-axis is used as the mirror surface, count the actual reference labels in the positioning area in the order from bottom to top and from left to right, and stop when it reaches half of the total number of actual reference labels in the system. The straight line represented by the abscissa is used as the right boundary of the area to be mirrored; the x-axis, the y-axis and the upper boundary of the positioning area are respectively used as the lower boundary, left boundary and upper boundary of the area to be mirrored; the actual reference label in the area to be mirrored And the virtual reference label is mapped to the y-axis according to the mirror imaging principle, and the mirrored area corresponding to the left becomes the mirrored reference label and the mirrored virtual label; the label in the mirrored area and the position coordinate of the label in the to-be-mirrored area before the mapping are symmetrical about the y-axis, image and thing respectively;

When taking the straight line where the right boundary of the positioning area is located as the mirror surface, count the actual reference labels in the positioning area in the order from bottom to top and from right to left, and stop when it reaches half of the total number of actual reference labels in the system. The straight line represented by the abscissa of the actual reference label is taken as the left border of the area to be mirrored; the x-axis, the upper border and the right border of the positioning area are taken as the lower border, upper border and right border of the area to be mirrored respectively; The actual reference label and virtual reference label in the mirror image are mapped to the right border of the positioning area according to the principle of mirror imaging, and the mirror area corresponding to the right becomes the mirror reference label and the mirror virtual label; The position coordinates are symmetrical about the right boundary of the positioning area, which are the image and the object respectively;

When the line where the upper boundary of the positioning area is located is the mirror surface, count the actual reference labels in the positioning area in the order from left to right and top to bottom, and stop when it reaches half of the total number of actual reference labels in the system. The straight line represented by the ordinate of the actual reference label is used as the lower boundary of the area to be mirrored; the y-axis, the upper boundary and the right boundary of the positioning area are respectively used as the left boundary, upper boundary and right boundary of the area to be mirrored. The actual reference label and the virtual reference label in the area to be mirrored are mapped to the corresponding mirror area above the upper boundary of the positioning area according to the principle of mirror imaging to become the mirror reference label and the mirror virtual label; the label in the mirror area and the area to be mirrored before mapping The position coordinates of the tags inside are symmetrical with respect to the upper boundary of the positioning area, which are the image and the object, respectively.

Further, in step S4, the absolute value of the difference between the two is multiplied by a constant coefficient of 0.55 as the threshold value when constructing a proximity map corresponding to reader i, each reader corresponds to a proximity map, including the entire positioning range and new The positioning area and the newly added mirroring area are divided into several small areas centered on the reference tag, and the RSSI value of the small area is taken as the RSSI value corresponding to the reference tag at the center of the area. After measuring the RSSI value of the tag, compare the value with the RSSI value of each area. If the difference between the two is within the calculated threshold range, the reader will mark the area as "1", and all readers in the system will be marked as "1". The corresponding adjacent maps are intersected, the small probability position of the label to be tested is excluded, and the reference label closest to the label to be tested is selected.

Further, in step S5, the first estimated position (x', y') of the label to be tested is calculated according to the final weight factor ω _i as follows:

Among them, (x _i , y _i ) is the position coordinate of the reference label retained after excluding the low probability position from the adjacent map, ω _i =ω _1i ×ω _2i , ω _1i represents the virtual reference retained after using the adjacent map to exclude the small probability position The difference of the signal strength value between the label and the label to be tested, ω _2i represents the density of the selected virtual reference label, and the magnitude of this value is positively correlated with the density.

Further, the difference ω _1i of the signal strength value between the virtual reference label and the label to be tested that is retained after excluding the low probability location using the proximity map is calculated as follows:

The density ω _2i of the selected virtual reference label is calculated as follows:

Among them, K represents the total number of readers in the system; _Sk (T _i ) represents the RSSI value of the i-th virtual reference tag corresponding to the k-th reader; _Sk (R) represents the collection of the k-th reader The RSSI value of the received label to be tested, na is the total number of reference labels selected by the small probability location exclusion method of the adjacent map; _pi is the number of regions directly connected to the region represented by the selected reference label _i and The ratio of the total number of areas to be tested; _nci is the number of areas connected to the area represented by the selected reference label i.

Further, in step S6, the threshold modification process is: calculating the absolute value of the difference between the RSSI value of the tag under test and the RSSI value of the reference tag with the largest difference, and multiplying the value by a new constant coefficient of 0.35 as a new threshold value.

Compared with the prior art, the present invention at least has the following beneficial effects:

The present invention is an RSSI indoor positioning method that introduces dynamic thresholds, which is based on VIRE and combines the principle of plane mirror imaging. First, add virtual reference labels with the same density as the central area in the positioning boundary area, and then use the positioning boundary line as a mirror surface to map various reference labels within nearly half of the positioning area to the other side of the boundary line to become mirror reference labels and mirror virtual labels, so that the border area indirectly becomes a "central area" surrounded by virtual labels. The mirroring algorithm with dynamic threshold can greatly improve the localization effect of VIRE algorithm, and can achieve better localization results than BVIRE.

Further, according to the RSSI of the actual reference tag read by the reader and the RSSI of the N*N distributed virtual reference tag calculated by the linear interpolation method mentioned in VIRE, a one-dimensional linear regression equation is established, which is the same as the one used in VIRE. The linear interpolation method is generally consistent in the idea of obtaining the RSSI value, and it is still assumed that the signal strength value follows a linear change. In addition, the use of a one-dimensional linear regression equation has the advantages of simple calculation and small error.

Further, the calculation method of the RSSI value of the virtual reference tag inserted in the boundary area is consistent with the linear interpolation method, so that the validity of the RSSI data in the system will be consistent, and based on a relatively simple idea and calculation method, it is possible to obtain the same value as VIRE. Reference data with consistent validity.

Further, through the study of the VIRE results, it can be found that the label to be tested located in the central area can achieve a better positioning effect than the boundary area, imitating this law, based on reasonable assumptions, through the system outside the boundary. By adding mirror labels, the boundary is located in the "center" of the new area, which is expected to improve the positioning accuracy at the boundary; the method of introducing mirror labels can enable the boundary area to be evenly covered by more virtual reference labels, which is the edge of the label to be tested at the boundary. The positioning provides more effective data information, which is expected to improve the positioning accuracy at the boundary.

Further, in order to make the reference position data selected by each tag to be tested more pertinent, the unified threshold of all tags to be tested in VIRE is changed to a dynamic threshold that differs depending on the tags to be tested, so as to obtain the most suitable value for the current tag to be tested. The reference tag of the measurement tag can effectively reduce the positioning error, and put forward a new idea for the future development of the application of indoor positioning using RFID technology.

Further, the setting method of the final weight factor is consistent with VIRE, which is divided into the product of two sub-factors, which can make more reasonable use of the "probability" relationship represented by the adjacent map. The weight factor setting method can improve the effectiveness of the final positioning result.

Further, since the ratio of the first set threshold selection is 0.55, it is more aimed at the label to be tested at the center, and it needs to be adjusted appropriately according to the first estimated position of the label to be tested. The adjustment method here is to first Determine whether the label to be tested is located in the boundary area, so as to decide whether to use the appropriate ratio of 0.35 for the boundary area as auxiliary data when setting the secondary threshold, which can be used for the first positioning. Correction to achieve an overall better positioning result.

To sum up, compared with VIRE, the present invention provides more information of the signal strength value of the tag as reference data by adding mirror tags on the outside of the boundary, which increases the base of the data volume; the dynamic threshold used can be more reasonable. The number of virtual reference labels introduced on the basis of VIRE is basically the same as that of BVIRE, and the use of dynamic thresholds enables this method to achieve a better overall positioning effect.

The technical solutions of the present invention will be further described in detail below through the accompanying drawings and embodiments.

Description of drawings

Fig. 1 is the layout diagram of the mirror algorithm system;

Fig. 2 is a boundary virtual reference label construction diagram;

Fig. 3 is the mapping process of boundary area reference label;

Fig. 4 is the small probability elimination method of the adjacent map;

Figure 5 is the distribution map of the points to be measured in the central area;

Figure 6 shows the positioning effect of different algorithms on the point to be measured in the central area;

Fig. 7 is the distribution map of the points to be measured in the boundary area;

Figure 8 shows the positioning effect of different algorithms on the point to be measured in the boundary area;

Figure 9 shows the positioning effects of different algorithms on the points to be measured in the overall area.

Detailed ways

The invention provides an RSSI indoor positioning method with dynamic threshold, which is based on VIRE and combines the principle of plane mirror imaging. First, add virtual reference labels with the same density as the central area in the positioning boundary area, and then use the positioning boundary line as a mirror surface to map various reference labels within nearly half of the positioning area to the other side of the boundary line to become mirror reference labels and mirror virtual labels, so that the border area indirectly becomes a "central area" surrounded by virtual labels. In order to optimize the most suitable reference label for each label to be tested, the algorithm has made changes in the threshold. The unified threshold used when constructing the proximity map is changed to a dynamic threshold that changes according to the different labels to be tested. The simulation results show that the mirror algorithm with dynamic threshold can greatly improve the localization effect of the VIRE algorithm, and its overall localization effect is better than that of BVIRE.

A kind of RSSI indoor positioning method that introduces dynamic threshold of the present invention, comprises the following steps:

S1. On the basis of the VIRE algorithm, establish a suitable one-dimensional linear regression equation according to the signal strength value and coordinate value of the actual reference label and the virtual reference label in the central area;

S2. Calculate the newly added boundary virtual reference labels on the boundary according to the one-dimensional linear regression equation established in step S1 (the distribution density of the virtual reference labels in the central part of VIRE is different, sometimes it needs to be added to the system, sometimes it is not necessary) signal strength value;

S3, use the linear interpolation method in the part between the four boundary lines and the outermost periphery of the central area to obtain the signal strength value of the virtual reference label added in this part;

S4. After the previous step, the area to be tested is now covered with various reference labels of the same density distribution (including the labels in the original VIRE and the newly added virtual reference labels). Then take the four boundary lines as mirror surfaces, and symmetrically map the labels in nearly half of the area to be measured on the inner side of the boundary line (here, the inner side refers to the direction of the area to be measured, and the outer side refers to the direction opposite to the area to be measured) to the outside of the boundary line, The actual reference label and the virtual reference label in the area to be tested are respectively mapped to the mirror reference label and the mirror virtual label in the mirror area;

S5. Calculate the signal strength value of the mirrored tag read by the reader in the system. The RSSI value of the tag read by some readers can be obtained directly according to "symmetrical equality", while others need to be obtained by regression equation method;

S6, according to the RSSI value of the tag to be tested read by the reader and the RSSI value of the above-mentioned various reference tags, find out the RSSI value of the reference tag with the largest difference with the RSSI value of the tag to be tested, and calculate the difference between the two. absolute value. Multiply this value by a constant coefficient of 0.55 as the threshold used when constructing a proximity map for the label to be tested;

S7, constructing a proximity map according to the threshold to screen out the reference labels suitable for the label to be measured, and to solve the weights of these reference labels, and then use the weight value and its coordinates to solve the position of the label to be measured;

S8. Determine whether the estimated position of the tag to be tested is located in the boundary area, and if so, modify the first threshold value. The modification process is: calculating the absolute difference between the RSSI value of the tag to be tested and the RSSI value of the reference tag with the largest difference. value, multiply this value by a new constant factor of 0.35 as the new threshold. The new thresholds are then used for steps such as building a proximity map, finding weight factors, and estimating the location of the label to be tested.

In order to make the purposes, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments These are some embodiments of the present invention, but not all embodiments. The components of the embodiments of the invention generally described and illustrated in the drawings herein may be arranged and designed in a variety of different configurations. Thus, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but is merely representative of selected embodiments of the invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

1. As shown in Figure 1, a typical VIRE algorithm layout has 16 actual reference labels distributed in the center of the 8*8 positioning area in a 4*4 structure, at the four corners of the positioning area, that is, the coordinates are (0,0 ), (0,8), (8,0) and (8,8) with 4 readers configured around. Virtual reference labels of N*N (at this time N=5) structure are uniformly distributed in a square formed by taking every four actual reference labels as the vertices, as shown in the right part indicated by the arrow in the figure. On this basis, an appropriate one-dimensional linear regression equation is established to solve the RSSI value of the border virtual reference tag read by each reader.

On the premise that the RSSI of the actual reference tag read by the reader and the RSSI of the virtual reference tag of N*N distribution calculated by the linear interpolation method mentioned in VIRE are known, the linear regression equation of one variable is established. Methods as below:

The basic expression of the linear equation in one variable is:

Among them, x is the abscissa of the position of the tag, and s is the RSSI value of the tag read by the reader;

和

值。Please refer to Figure 2, taking the four actual reference labels T ₁ , T ₂ , T ₃ , T ₄ and multiple virtual reference labels such as t1, t2, t3, t4, t5, t6 as an example, if the reader in the system The number of is n, the RSSI values of these four actual reference tags and multiple virtual reference tags on reader i are S _i1 , S _i2 , S _i3 , S _i4 and s _i1 , s _i2 , s _i3 , s respectively _i4 , s _i5 , s _i6 , where 1≤i≤n, and the abscissas of their positions in the system are X ₁ , X ₂ , X ₃ , X ₄ and x ₁ , x ₂ , x ₃ , x respectively ₄ , x ₅ , x ₆ . Then, in relation to formula (1-1), the relevant RSSI values and the abscissa data sets (X ₁ , S _i1 ), (X ₂ , S _i2 ), (X ₃ , S _i3 ), (X ₄ , S _i4 ) and (x ₁ ,s _i1 ),(x ₂ ,s _i2 ),(x ₃ ,s _i3 ),(x ₄ ,s _i4 ),(x ₅ ,s _i5 ),(x ₆ ,s _i6 ), etc. respectively Substitute into equations (1-2) and (1-3) to obtain a set of

and

value.

Then, according to the formula (1-1) and the abscissa data of the boundary virtual reference tags T ₁₁ and T ₁₂ , their RSSI values on the reader i can be obtained.

Using this method, the RSSI values read by the readers of the virtual reference tags on the left and right borders of the positioning area can be obtained. As for the RSSI values of the virtual reference tags on the upper and lower boundaries of the positioning area, it can be obtained by a method similar to the above, just replace x in the above three formulas with y.

2. Then use the linear interpolation method to obtain the RSSI value of the newly added virtual reference label in the part between the four boundary lines and the outermost periphery of the central area.

The linear interpolation formula is as follows:

In the horizontal direction:

In the vertical direction:

0≤p＝i％n≤n-1，0≤q＝j％n≤n-1。In the above two formulas, _Sk (T _i,j ) represents the RSSI value of the virtual reference tag at (i,j) read by the kth reader, where,

0≤p=i%n≤n-1, 0≤q=j%n≤n-1.

After linear interpolation is used in the boundary area, the boundary part and the central part of the positioning area are also distributed with the same density of virtual reference labels. For a specific illustration, please refer to the first three rows of labels (partially) indicated by the lower arrows in FIG. 1 , wherein the third row is a boundary virtual label on the lower boundary.

3. For the classic layout in Figure 1, the system has a total of four readers, all of which are distributed on the boundary line of the positioning area. The upper, lower, left, and right boundary lines are respectively used as mirror surfaces. Through the mapping function, the mirror label outside the boundary is added to the system.

The specific method of joining is as follows:

① Taking the x-axis as the mirror surface, the actual reference labels and virtual reference labels in the area enclosed by the x-axis, y-axis, y=3 and x=8 are mapped to the x-axis, y-axis, y=-3 according to the principle of mirror imaging The area bounded by x=8 becomes the mirror reference label and the mirror virtual label;

② With the y-axis as the mirror surface, the actual reference labels and virtual reference labels in the area enclosed by the x-axis, y-axis, x=3 and y=8 are mapped to the x-axis, y-axis, x=-3 according to the principle of mirror imaging and y=8 as the boundary area becomes the mirror reference label and the mirror virtual label;

③ Taking the line x=8 as the mirror surface, the actual reference label and virtual reference label in the area enclosed by the x-axis, x=5, x=8 and y=8 are mapped to the x-axis, x= 8. The area bounded by x=11 and y=8 becomes the mirror reference label and the mirror virtual label;

④ Taking the line y=8 as the mirror surface, the actual reference label and virtual reference label in the area enclosed by the y-axis, x=8, y=5 and y=8 are mapped to the y-axis, x= 8. The area bounded by y=8 and y=11 becomes the mirror reference label and the mirror virtual label.

After the mapping of the above four boundary lines, the overall layout of the system is shown in Figure 3 (due to the excessive number of labels, only part of the figure is shown).

Here, only the left border is used as an example to describe how to calculate the RSSI value of the mirrored label in the mirrored area:

For the left border of the area, due to the symmetrical distribution of the labels on both sides, the various angles and distance information between them and the border line are symmetrical and the same. Therefore, the No. 0 reader and No. 1 reader on the border line read The obtained RSSI values of the mirrored reference label and the mirrored virtual label on the left side of the boundary should be correspondingly equal to the RSSI values of the actual reference label and the virtual reference label on the right side of the boundary.

For readers No. 2 and No. 3 in the figure, the position and angle information of the "object" and "image" on the left border no longer have a symmetric relationship. Therefore, they read the left border. The RSSI value of the mirror reference label and the mirror virtual label on the side cannot be obtained directly according to the "symmetrical equality" as above, but another method is required.

Here, for readers No. 2 and No. 3, the RSSI value of the actual reference tag on the right side of the left border and the virtual reference tag on the right can be calculated by using the regression equation method described above to obtain the mirror reference tag on the left side of the left border and the mirror virtual reference tag. The tag's RSSI value on the corresponding reader.

4. So far, the mirroring algorithm has obtained the RSSI values of various tags in the system on each reader. Next, according to the RSSI value of the tag to be tested read by the reader i, find out the corresponding reader and the tag to be tested. The RSSI value with the largest signal strength difference is obtained, and the absolute value of the difference between the two is obtained, and then the value is multiplied by a constant coefficient of 0.55 as the threshold value when constructing the proximity map corresponding to reader i.

The process of building a proximity map is shown in Figure 4: each reader corresponds to a proximity map, which includes the entire positioning range, and divides the positioning area into multiple reference tags (including actual reference tags and virtual reference tags) as For the small area in the center, its RSSI value is taken as the RSSI value corresponding to the reference label at the center of the area. When the reader reads the RSSI value of the tag to be tested, it compares it with the RSSI value of each area. If the difference between the two is within the calculated threshold range, the reader marks the area as "1" (as shown by the black square in the picture). Finally, the intersection of the adjacent maps corresponding to all readers in the system is taken, and the small probability position of the tag to be tested is excluded, so as to select the reference tag closest to the tag to be tested.

5. After obtaining the reference labels closest to the label to be tested, it is necessary to solve their weight values. Here, the weight factor is still the product of the two sub-factors in VIRE. The specific formula is as follows:

ω _1i represents the difference in signal strength values between the virtual reference tag and the tag to be tested that remain after excluding small probability locations using the proximity map. In formula (1-6), K represents the total number of readers in the system; S _k ( T _i ) represents the RSSI value of the i-th virtual reference tag corresponding to the k-th reader; _Sk (R) represents the RSSI value of the tag to be tested collected by the k-th reader.

ω _2i represents the density of the selected virtual reference label, and the size of this value is positively correlated with the density. In formula (1-7), n _a is the total number of reference labels selected by the small probability location exclusion method of the adjacent map. pi is the ratio of the number of areas directly connected to the area represented by the selected reference label _i to the total number of areas to be tested; n _ci is the number of areas connected to the area represented by the selected reference label i .

The final weight factor formula is as follows:

ω _i =ω _1i ×ω _2i (1-8)

According to the weight value in formula (1-8), the estimated position of the label to be tested can be calculated as

Among them, (x _i , y _i ) are the position coordinates of the reference label retained after excluding small probability positions from the adjacent map.

6. According to the estimated position of the label to be tested calculated in the previous step, determine whether it is located in the boundary area, that is, the part between the four boundary lines and the outermost periphery of the central area. If it is located in this area, you need to re-adjust the threshold for two The second positioning is performed, and the first positioning result is corrected.

The new threshold setting method at this time is: modify the constant coefficient in the fourth step to 0.35, and then perform the subsequent steps of constructing a proximity map, determining the weight factor of the reference label, and secondary positioning.

Simulation and Analysis

Simulation environment: Select the area to be measured as shown in Figure 1, and take 1000 random points in the central area (as shown in Figure 5) as the points to be measured for simulation. The simulation parameters n, N, th, k, etc. are all selected to make the algorithms such as LANDMARC, VIRE and BVIRE have better positioning effects.

Figure 6 shows the cumulative distribution of errors after locating these 1000 random points to be measured using four different algorithms. The simulation diagram shows that the mirroring algorithm improves the positioning effect of VIRE at the center a lot, and at the same time, it can achieve better positioning than BVIRE. Better positioning effect.

Then take 1000 random points in the boundary area (as shown in Figure 7) as the points to be measured for simulation. Figure 8 shows the cumulative distribution of errors after using four different algorithms to locate the points to be measured. The simulation diagram shows that the mirroring algorithm also improves the positioning effect of VIRE at the boundary, and can achieve the same level of positioning as BVIRE. Effect.

Finally, the central area and the boundary area are integrated, and the positioning effect of different algorithms on the entire area to be measured is analyzed. Considering that the main activity range of indoor objects is still the central area, when selecting the points to be measured, a total of 1000 random numbers are selected as the points to be measured, of which 900 are located in the central area and the other 100 are located in the boundary area. The final simulation result is shown in Figure 9. It can be seen from the figure that the mirroring algorithm significantly improves the positioning effect of VIRE, and its overall positioning effect is better than that of BVIRE.

The above content is only to illustrate the technical idea of the present invention, and cannot limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention all fall within the scope of the claims of the present invention. within the scope of protection.

Claims (6)

1. An RSSI indoor positioning method introducing dynamic threshold is characterized in that a VIRE algorithm is utilized to establish a unary linear regression equation of signal intensity values and coordinate values of an actual reference label and a virtual reference label, the signal intensity value of the virtual reference label in a newly added boundary area is obtained through calculation, and the virtual reference label with the same density distribution as that of a central area is added in a positioning boundary area; then, taking the positioning boundary line as a mirror surface, mapping the reference label in the positioning area to the other side of the boundary line to form a mirror image reference label and a mirror image virtual label, and indirectly forming the boundary area into a central area surrounded by the virtual label at four circles; according to the RSSI value of the tag to be detected and the RSSI value of the reference tag read by the reader, finding out the RSSI value of the reference tag which has the maximum difference with the RSSI value of the tag to be detected to obtain the absolute value of the difference between the RSSI value and the RSSI value of the reference tag, determining the threshold value used when the tag to be detected constructs the adjacent map, and modifying the uniform threshold value used when the adjacent map is constructed into a dynamic threshold value which is changed according to the difference of the tag to be detected; then, constructing an adjacent map according to the dynamic threshold value to screen out a reference label suitable for the label to be detected, solving the position of the label to be detected according to the weight value and the coordinate of the reference label, and completing the positioning of the target in the whole area to be detected, wherein the method comprises the following steps:

s1, configuring 4 readers at four corners of the positioning area, uniformly distributing virtual reference labels with an N-N structure in a square formed by taking every four actual reference labels as vertex angles, and establishing a unitary linear regression equation according to the RSSI of the actual reference labels read by the readers and the RSSI of the N-N distributed virtual reference labels calculated by a linear interpolation method mentioned in VIRE;

s2, calculating the signal strength value of each newly added boundary virtual reference label on the boundary according to the unitary linear regression equation established in step S1, obtaining the RSSI value of the newly added virtual reference label by using a linear interpolation method at the portion between the four boundary lines and the outermost periphery of the central region, wherein the linear interpolation formula in the horizontal direction is as follows:

the linear interpolation formula in the vertical direction is as follows:

wherein S is_k(T_i,j) Represents the RSSI value of the virtual reference tag at (i, j) read by the kth reader,

0≤p＝i％n≤n-1，0≤q＝j％n≤n-1；

s3, taking four boundary lines as mirror surfaces, symmetrically mapping labels in approximately half of a region to be detected on the inner side of the boundary lines to the outer sides of the boundary lines, respectively mapping an actual reference label and a virtual reference label in the region to be detected to a mirror image reference label and a mirror image virtual label in a mirror image region, and calculating the RSSI value of the mirror image label read by a reader in the system;

s4, according to the RSSI value of the to-be-detected label read by the reader and the RSSI value of the reference label, finding the RSSI value of the reference label which has the maximum difference with the RSSI value of the to-be-detected label, calculating the absolute value of the difference between the RSSI value and the RSSI value of the reference label, determining the threshold value used when the to-be-detected label constructs the adjacent map, multiplying the absolute value of the difference between the RSSI value and the threshold value by a constant coefficient 0.55 to be used as the threshold value when the adjacent map corresponding to the reader i is constructed, wherein each reader corresponds to one adjacent map and comprises the whole positioning range and a newly-added mirror range, dividing the positioning range and the newly-added mirror range into a plurality of small areas taking the reference label as the center, taking the RSSI value of the small area as the RSSI value corresponding to the reference label at the center of the area, comparing the RSSI value with the RSSI value of each area after the reader reads the RSSI value of the to-detected label, if, taking intersection of adjacent maps corresponding to all readers in the system, excluding the small probability position of the tag to be detected, and selecting a reference tag close to the tag to be detected;

s5, constructing an adjacent map according to a threshold value, screening out reference labels suitable for the labels to be tested, solving the weights of the reference labels, and then solving the estimated positions (x ', y') of the labels to be tested by using the weight values and the coordinates thereof;

s6, if the first estimated position (x ', y') of the to-be-detected label is located in the boundary area, multiplying the absolute value in the step S4 by 0.35 to serve as a secondary threshold used when the to-be-detected label constructs the adjacent map, and then using the new threshold to construct the adjacent map, and solving the weighting factor and the estimated position of the to-be-detected label; if the first estimated position (x ', y') of the tag to be detected is located in the central area, the original threshold value and the positioning result are not changed.

2. The RSSI indoor positioning method with introduced dynamic threshold as claimed in claim 1, wherein in step S1, the unary linear equation is as follows:

wherein,

x is the abscissa of the position of the tag, S is the RSSI value read by the reader to the tag, the number of the readers in the system is set as n, and the RSSI values of the four actual reference tags and the virtual reference tags on the reader i are respectively set as S_i1，S_i2，S_i3，S_i4And s_i1，s_i2，s_i3，s_i4，s_i5，s_i6I is more than or equal to 1 and less than or equal to n, and the abscissa of the position in the system is X₁，X₂，X₃，X₄And x₁，x₂，x₃，x₄，x₅，x₆RSSI value and abscissa data set are (X)₁,S_i1)，(X₂,S_i2)，(X₃,S_i3)，(X₄,S_i4) And (x)₁,s_i1)，(x₂,s_i2)，(x₃,s_i3)，(x₄,s_i4)，(x₅,s_i5)，(x₆,s_i6)。

3. The RSSI indoor positioning method as claimed in claim 1, wherein in step S3, four boundary lines, i.e. upper, lower, left and right, are used as mirror surfaces respectively, and mirror labels outside the boundary are added to the system through mapping, specifically as follows:

counting the actual reference labels in the positioning area from left to right and from bottom to top in sequence when the x axis is taken as a mirror surface until the actual reference labels reach half of the total number of the actual reference labels in the system, and taking a straight line represented by a vertical coordinate of the corresponding actual reference label at the moment as an upper boundary of the area to be mirrored; respectively taking the x axis, the y axis and the right boundary of the positioning area as the lower boundary, the left boundary and the right boundary of the area to be mirrored; mapping the actual reference label and the virtual reference label in the area to be mirrored to a corresponding mirror area below an x axis according to a mirror imaging principle to form a mirror reference label and a mirror virtual label; the position coordinates of the label in the mirror image area and the label in the area to be mirrored before mapping are symmetrical about an x axis and are an image and an object respectively;

counting the actual reference labels in the positioning area from bottom to top in the sequence from left to right when the y axis is taken as the mirror surface until the actual reference labels reach half of the total number of the actual reference labels in the system, and taking the straight line represented by the abscissa of the corresponding actual reference label at the moment as the right boundary of the area to be mirrored; respectively taking the upper boundaries of the x axis, the y axis and the positioning area as the lower boundary, the left boundary and the upper boundary of the area to be mirrored; mapping the actual reference label and the virtual reference label in the region to be mirrored to a y axis according to a mirror imaging principle, and enabling the mirror region corresponding to the left to be a mirror image reference label and a mirror image virtual label; the position coordinates of the label in the mirror image area and the label in the area to be mirrored before mapping are symmetrical about the y axis and are an image and an object respectively;

when the straight line of the right boundary of the positioning area is taken as the mirror surface, counting the actual reference labels in the positioning area from bottom to top and from right to left until the number of the actual reference labels in the system is half of the total number, and taking the straight line represented by the abscissa of the corresponding actual reference label at the moment as the left boundary of the area to be mirrored; respectively taking the x axis, the upper boundary and the right boundary of the positioning area as the lower boundary, the upper boundary and the right boundary of the area to be mirrored; mapping the actual reference label and the virtual reference label in the area to be mirrored to the right boundary of the positioning area according to the mirror imaging principle, and enabling the mirror area corresponding to the right to be the mirror image reference label and the mirror image virtual label; the position coordinates of the label in the mirror image area and the label in the area to be mirrored before mapping are symmetrical relative to the right boundary of the positioning area and are an image and an object respectively;

when the straight line where the upper boundary of the positioning area is located is taken as the mirror surface, counting the actual reference labels in the positioning area from left to right and from top to bottom in sequence until the number of the actual reference labels in the system is half of the total number, and taking the straight line represented by the vertical coordinate of the corresponding actual reference label at the moment as the lower boundary of the area to be mirrored; respectively taking the upper boundary and the right boundary of the y-axis positioning area as the left boundary, the upper boundary and the right boundary of the area to be mirrored; mapping the actual reference label and the virtual reference label in the area to be mirrored to a corresponding mirror area above the upper boundary of the positioning area according to a mirror imaging principle to form a mirror image reference label and a mirror image virtual label; the position coordinates of the label in the mirror image area and the label in the area to be mirror imaged before mapping are symmetrical about the upper boundary of the positioning area and are an image and an object respectively.

4. The RSSI indoor positioning method with the dynamic threshold introduced as claimed in claim 1, wherein in step S5, the final weighting factor ω is determined according to_iCalculating to obtain the first estimated position (x ', y') of the to-be-detected label as follows:

wherein (x)_i,y_i) Position coordinates, omega, of reference labels reserved for excluding small probability positions via a proximity map_i＝ω_1i×ω_2i，ω_1iRepresenting the difference in signal strength values, ω, between the virtual reference tag remaining after excluding a small probability location using a proximity map and the tag under test_2iThe density of the selected virtual reference label is shown, and the magnitude of the value is in positive correlation with the density.

5. The RSSI indoor positioning method of claim 4, wherein the difference ω in signal strength values between the virtual reference tag and the tag under test that remain after excluding the small probability location using the proximity map is characterized by_1iThe calculation is as follows:

density omega of selected virtual reference tags_2iThe calculation is as follows:

wherein K represents the total number of readers in the system; s_k(T_i) The RSSI value of the corresponding ith virtual reference label on the kth reader is represented; s_k(R) represents the RSSI value of the to-be-detected label collected by the kth reader, n_aThe total number of the reference labels selected by the small probability position elimination method of the adjacent map; p is a radical of_iIs the ratio of the number of the areas directly connected with the area represented by the selected reference label i to the total number of the areas to be detected; n is_ciIs the number of regions linked to the region represented by the selected reference label i.

6. The RSSI indoor positioning method with introduced dynamic threshold as claimed in claim 1, wherein in step S6, the threshold is modified by: and calculating the absolute value of the difference between the RSSI value of the tag to be detected and the RSSI value of the reference tag which has the largest difference with the RSSI value, and multiplying the absolute value by a new constant coefficient 0.35 to serve as a new threshold value.

Images