A distance estimation method and system for suppressing complex indoor RSSI fluctuations

By performing data conditioning, time domain and frequency domain optimization on RSSI signals, and combining logarithmic distance path attenuation model, the problem of RSSI fluctuations affecting indoor positioning in complex indoor environments is solved, and positioning accuracy and stability are improved.

CN113296052BActive Publication Date: 2024-11-15HUNAN UNIV OF SCI & TECH +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110553292.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-05-20
Publication Date
2024-11-15
Estimated Expiration
2041-05-20

AI Technical Summary

Technical Problem

The fluctuations in RSSI signal intensity caused by complex indoor environments affect the accuracy and stability of indoor positioning algorithms. The prior art is difficult to effectively suppress such fluctuations, resulting in low positioning accuracy.

Method used

Through RSSI data conditioning, time domain optimization, frequency domain optimization and RSSI screening steps, combined with the logarithmic distance path attenuation model, smooth filtering and low-frequency partial extraction of the RSSI signal are realized, and stable RSSI values ​​are calculated to estimate the distance.

Benefits of technology

It effectively reduces the random error of the positioning algorithm in the indoor environment and improves the indoor positioning accuracy and stability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113296052B_ABST
    Figure CN113296052B_ABST
Patent Text Reader

Abstract

A distance estimation method and system for suppressing complex indoor RSSI fluctuations, comprising: an RSSI data conditioning step, receiving random raw received signal strength data D sequentially input within a period of time RSSI , construct a receiving window for multiple data, input one data at a time, select the median in the window, and sequentially calculate the absolute difference between two adjacent data and accumulate them to obtain the absolute difference cumulative value; the time domain optimization step, according to the size of the absolute difference cumulative value and the median ratio, the original data D RSSI Perform smoothing filtering to obtain data D gus ; Frequency domain optimization step, the data D gus Transform into frequency domain data, extract the low-frequency part of RSSI in the frequency domain, and then transform into time domain data D LF ; RSSI screening steps, statistics D LF The number of occurrences of RSSI data in different intervals is calculated, and the data with a number of occurrences below the screening threshold is removed, and the geometric mean RD of the remaining data is calculated. RSSI That is the estimated RSSI value; in the distance estimation step, the distance is estimated based on the mathematical relationship between the RSSI and the logarithmic distance path attenuation model between the paths.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of digital signal processing, and in particular to an indoor positioning algorithm using RSSI signal strength values ​​as wireless short-range communication. Background Art

[0002] With the increasing number of large buildings and the rapid development of Internet of Things technology, the demand for indoor location services based on short-range wireless communication technologies such as Wi-Fi, ZigBee, UWB, and Bluetooth is becoming an important indoor positioning service to meet people's pursuit of convenient life. The low-cost and low-complexity positioning algorithm based on RSSI signal strength is suitable for indoor positioning requirements in some daily life occasions. However, the complex indoor environment causes the RSSI intensity to fluctuate greatly, which seriously affects the accuracy and stability of the indoor positioning algorithm based on RSSI as the ranging. Summary of the invention

[0003] The purpose of the present invention is to provide a distance estimation method and system for suppressing complex indoor RSSI fluctuations, so as to reduce the random error caused by the influence of the indoor environment and improve the accuracy of the indoor positioning algorithm.

[0004] The present invention provides a distance estimation method for suppressing complex indoor RSSI fluctuations, comprising: an RSSI data conditioning step, receiving random raw received signal strength data D sequentially input within a period of time; RSSI , construct a receiving window for multiple data, input one data at a time, select the median in the window, and sequentially calculate the absolute difference between two adjacent data and accumulate them to obtain the absolute difference cumulative value; the time domain optimization step, according to the size of the absolute difference cumulative value and the median ratio, the original data D RSSI Perform smoothing filtering to obtain data D gus ; Frequency domain optimization step, the data D gus Transform into frequency domain data, extract the low-frequency part of RSSI in the frequency domain, and then transform into time domain data D LF ; RSSI screening steps, statistics D LF The number of occurrences of RSSI data in different intervals is calculated, and the data with a number of occurrences below the screening threshold is removed, and the geometric mean RD of the remaining data is calculated. RSSI That is the estimated RSSI value; in the distance estimation step, the distance is estimated based on the mathematical relationship between the RSSI and the logarithmic distance path attenuation model between the paths.

[0005] Furthermore, in the RSSI data conditioning step, the number of data in the receiving window is 5, and the absolute difference accumulation value A n The calculation method is as follows: Based on the original RSSI sequence data D_RSSI n , the nth absolute difference is: The cumulative sum of absolute differences is: Where E -1 =E3,E -2 =E2, n∈N+; median M n The calculation method is as follows: Based on the current original RSSI data D n , from 5 consecutive raw RSSI data D n-2 , D n-1 , D n , D n+1 , D n+2 Select the middle value in the number.

[0006] Furthermore, in the time domain optimization step, the original data D RSSI The smoothing filter processing includes: according to S n =A n / M n The value of determines the smoothness of the filter window, and is used as the basis for selecting the size of the Gaussian function smoothing scale, thereby adjusting the weighting coefficient of the filter and realizing the adaptive adjustment of the filter scale parameter. The rules for adaptive adjustment of the standard deviation σ are as follows:

[0007]

[0008] Furthermore, in the frequency domain optimization step, the data D gus Transformed into frequency domain data, extracting the low-frequency part of RSSI in the frequency domain includes: gus After Fourier transformation into frequency domain amplitude data, the data group with the largest amplitude is selected as the low-frequency range. The maximum frequency of this group is set as the low-frequency dividing point. All data below the low-frequency dividing point are taken as the low-frequency range, and the amplitude of the high-frequency part is set to 0.

[0009] Furthermore, the method for setting the low-frequency demarcation point includes: dividing the frequency domain amplitude data into 4 groups with reference to the K-means clustering algorithm, and the steps of the clustering algorithm are as follows:

[0010] S1: Set the number of groups k = 4, and select the amplitudes of 4 frequency points at random as the center points;

[0011] S2: Calculate the numerical difference between the amplitude of all frequency points and the center point, select the center point with the smallest numerical difference as the same group, and divide the amplitude of all frequency points into k groups;

[0012] S3: Calculate the average value of the data in the group, use the average value as the new center point of each group, and return to S2 until it is iterated 5 times;

[0013] After the grouping is completed, the amplitude sums of the four groups are compared. The group with the largest sum is the target group. The highest amplitude frequency point in this group is taken as the low-frequency dividing point, and the frequency below the dividing point is the low-frequency range.

[0014] Furthermore, in the RSSI screening step, the step size of the partition is: (max(D LF )-min(D LF )) / 5, the interval with the largest number of RSSI occurrences is taken as the basic unit, and the screening threshold is 1 / 2 of the number of RSSI occurrences in the basic unit.

[0015] Furthermore, in the distance estimation step, the logarithmic distance path attenuation model is: RSSI=RSSI0+10μlg(d / d0)+ε, wherein RSSI0 is the RSSI value at the unit distance d0; ε is the environmental error coefficient; and μ is the path attenuation index.

[0016] The present invention also provides a distance estimation system for suppressing complex indoor RSSI fluctuations, comprising: an RSSI data conditioning module for receiving random raw received signal strength data D sequentially input within a period of time RSSI , construct a receiving window for multiple data, input one data at a time, select the median in the window, and sequentially calculate the absolute difference between two adjacent data and accumulate them to obtain the absolute difference cumulative value; the time domain optimization module is used to optimize the original data D according to the ratio of the absolute difference cumulative value to the median. RSSI Perform smoothing filtering to obtain data D gus ; Frequency domain optimization module, used to convert data D gus Transform into frequency domain data, extract the low-frequency part of RSSI in the frequency domain, and then transform into time domain data D LF ; RSSI screening module, used to count D LF The number of occurrences of RSSI data in different intervals is calculated, and the data with a number of occurrences below the screening threshold is removed, and the geometric mean RD of the remaining data is calculated. RSSI That is the estimated RSSI value; the distance estimation module is used to estimate the distance according to the mathematical relationship between RSSI and the logarithmic distance path attenuation model between the paths.

[0017] The present invention provides a distance estimation method and system for suppressing complex indoor RSSI fluctuations. The method combines and optimizes RSSI reception values ​​in the time domain and frequency domain, obtains estimated RSSI values ​​after screening and optimization, and finally outputs distance estimation values ​​using a logarithmic distance attenuation model. This can reduce random errors caused by the influence of the indoor environment and improve the accuracy of the indoor positioning algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1The present invention is a block diagram showing the principle of a distance calculation method for suppressing RSSI fluctuations in a complex indoor environment according to an embodiment of the present invention.

[0019] Figure 2 for Figure 1 Block diagram of the Gaussian filter distance in the time domain optimization step of the distance calculation method shown. DETAILED DESCRIPTION

[0020] like Figure 1 As shown, in an embodiment of the present invention, a distance estimation method for suppressing complex indoor RSSI fluctuations includes five steps: RSSI data conditioning, time domain optimization, frequency domain optimization, RSSI screening, and distance estimation.

[0021] The RSSI data conditioning step includes: receiving random raw received signal strength data D sequentially input within a period of time RSSI , construct a receiving window of multiple data, input one data at a time, select the median in the window, and sequentially calculate the absolute difference between two adjacent data and accumulate them to obtain the absolute difference cumulative value. In this embodiment, the number of data in the receiving window is 5, and the absolute difference cumulative value A n The calculation method is as follows: Based on the original RSSI sequence data D_RSSI n , the nth absolute difference is: The cumulative sum of absolute differences is: Where E -1 =E3,E -2 =E2, n∈N+; median M n The calculation method is as follows: Based on the current original RSSI data D n , from 5 consecutive raw RSSI data D n-2 , D n-1 , D n , D n+1 , D n+2 Select the middle value in the number.

[0022] The time domain optimization step includes: according to the size of the ratio of the absolute difference accumulation value to the median value, the original data D RSSI Perform smoothing filtering to obtain data D gus Among them, for the original data D RSSI The smoothing filter processing includes: according to S n =A n / M n The value of determines the smoothness of the filter window, and takes it as the basis for selecting the size of the Gaussian function smoothing scale, thereby adjusting the weighting coefficients [H1, H2, H3, H4, H5] of the filter to achieve adaptive adjustment of the filter scale parameters. The rules for adaptive adjustment of the standard deviation σ are as follows:

[0023] Since H1=H5, H2=H4, considering the calculation speed, we can design Figure 2 The filtering calculation method shown.

[0024] The frequency domain optimization steps include: gus Transform into frequency domain data, extract the low-frequency part of RSSI in the frequency domain, and then transform into time domain data D LF Wherein, the data D gus Transformed into frequency domain data, extracting the low-frequency part of RSSI in the frequency domain includes: gus After Fourier transformation into frequency domain amplitude data, the data group with the largest amplitude is selected as the low-frequency range. The maximum frequency of this group is set as the low-frequency dividing point. All data below the low-frequency dividing point is taken as the low-frequency range, and the amplitude of the high-frequency part is set to 0, that is, the high-frequency component is deleted.

[0025] The method for setting the low-frequency demarcation point includes: dividing the frequency domain amplitude data into four groups with reference to the K-means clustering algorithm, and the steps of the clustering algorithm are as follows:

[0026] S1: Set the number of groups k = 4, and select the amplitudes of 4 frequency points at random as the center points;

[0027] S2: Calculate the numerical difference between the amplitude of all frequency points and the center point, select the center point with the smallest numerical difference as the same group, and divide the amplitude of all frequency points into k groups;

[0028] S3: Calculate the average value of the data in the group, use the average value as the new center point of each group, and return to S2 until it is iterated 5 times;

[0029] After the grouping is completed, the sum of the amplitudes of the four groups is compared. The group with the largest sum is the target group, and the highest amplitude frequency point in this group is used as the low-frequency demarcation point, and the frequency below the demarcation point is the low-frequency range. This setting method can adaptively adjust the low-frequency range according to the input data, retain the characteristics of the original number as much as possible, achieve low-computational complexity low-frequency filtering, and achieve fast convergence.

[0030] RSSI screening steps include: Statistics D LF The number of occurrences of RSSI data in different intervals is calculated, and the data with a number of occurrences below the screening threshold is removed, and the geometric mean RD of the remaining data is calculated. RSSI This is the estimated RSSI value. In this way, by partitioning the RSSI value output by the previous module, the number of RSSI occurrences in each interval is counted, and the RSSI value with a smaller number of occurrences is removed. The step size of the partition is: (max(D LF )-min(D LF)) / 5, the interval with the largest number of RSSI occurrences is taken as the basic unit, and the screening threshold is 1 / 2 of the number of RSSI occurrences in the basic unit.

[0031] The distance estimation step includes: estimating the distance according to the mathematical relationship between RSSI and the logarithmic distance path attenuation model between the paths. The logarithmic distance path attenuation model is: RSSI = RSSI0 + 10μlg (d / d0) + ε, where RSSI0 is the RSSI value at the unit distance d0; ε is the environmental error coefficient; μ is the path attenuation index. In this embodiment, the parameters d0 can be set to 1, ε=0, μ=2.46, so the estimated distance is: d=10(RSSI-75) / 24.6.

[0032] The embodiment of the present invention also provides a distance estimation system for suppressing complex indoor RSSI fluctuations, including an RSSI data conditioning module, a time domain optimization module, a frequency domain optimization module, an RSSI screening module, and a distance estimation module. The RSSI data conditioning module is used to receive random raw received signal strength data D sequentially input within a period of time. RSSI , construct a receiving window for multiple data, input one data at a time, select the median in the window, and sequentially calculate the absolute difference between two adjacent data and accumulate them to obtain the absolute difference cumulative value. The time domain optimization module is used to optimize the original data D according to the ratio of the absolute difference cumulative value to the median. RSSI Perform smoothing filtering to obtain data D gus The frequency domain optimization module is used to transform the data D gus Transform into frequency domain data, extract the low-frequency part of RSSI in the frequency domain, and then transform into time domain data D LF RSSI filtering module, used to count D LF The number of occurrences of RSSI data in different intervals is calculated, and the data with a number of occurrences below the screening threshold is removed, and the geometric mean RD of the remaining data is calculated. RSSI The distance estimation module is used to estimate the distance according to the mathematical relationship between RSSI and the logarithmic distance path attenuation model between the paths.

[0033] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A distance estimation method for suppressing complex indoor RSSI fluctuations, characterized in that: include: RSSI data conditioning step, receiving random raw received signal strength data D input sequentially over a period of time RSSI , construct a receiving window for multiple data, input one data at a time, select the median in the window, and sequentially calculate the absolute difference between two adjacent data and accumulate them to obtain the absolute difference cumulative value; In the time domain optimization step, the original data D is optimized according to the ratio of the absolute difference accumulation value to the median value. RSSI Perform smoothing filtering to obtain data D gus ; Frequency domain optimization step, the data D gus Transform into frequency domain data, extract the low-frequency part of RSSI in the frequency domain, and then transform into time domain data D LF ; RSSI Screening step, statistics D LF The number of occurrences of RSSI data in different intervals is calculated, and the data with a number of occurrences below the screening threshold is removed, and the geometric mean RD of the remaining data is calculated. RSSI That is the estimated RSSI value; Distance estimation step, estimating the distance based on the mathematical relationship between RSSI and the logarithmic distance path attenuation model between the paths; In the frequency domain optimization step, the data D gus Transformed into frequency domain data, extracting the low-frequency part of RSSI in the frequency domain includes: gus After Fourier transformation, the frequency domain amplitude data is converted into the data. The data group with the largest amplitude is selected as the low-frequency range. The maximum frequency of this group is set as the low-frequency cutoff point. All data below the low-frequency cutoff point are taken as the low-frequency range. The amplitude of the high-frequency part is set to 0. The method for setting the low-frequency demarcation point includes: dividing the frequency domain amplitude data into four groups with reference to the K-means clustering algorithm, and the steps of the clustering algorithm are as follows: S1: Set the number of groups k = 4, and select the amplitudes of 4 frequency points at random as the center points; S2: Calculate the numerical difference between the amplitude of all frequency points and the center point, select the center point with the smallest numerical difference as the same group, and divide the amplitude of all frequency points into k groups; S3: Calculate the average value of the data in the group, use the average value as the new center point of each group, and return to S2 until it is iterated 5 times; After the grouping is completed, the sum of the amplitudes of the four groups is compared. The group with the largest sum is the target group. The highest amplitude frequency point in this group is used as the low-frequency demarcation point, and the frequency below the demarcation point is in the low-frequency range. In the distance estimation step, the logarithmic distance path attenuation model is: RSSI=RSSI0+10μlg(d / d0)+ε, where RSSI0 is the RSSI value at unit distance d0; ε is the environmental error coefficient; μ is the path attenuation index; and d is the estimated distance.

2. The distance estimation method according to claim 1, characterized in that: In the RSSI data conditioning step, the number of data in the receiving window is 5, and the absolute difference accumulation value A n The calculation method is as follows: Based on the original RSSI sequence data D_RSSI n , the nth absolute difference is: The cumulative sum of absolute differences is: Where E -1 =E3,E -2 =E2, n∈N+; median M n The calculation method is as follows: Based on the current original RSSI data D n , from 5 consecutive raw RSSI data D n-2 , D n-1 , D n , D n+1 , D n+2 Select the middle value in the .

3. The distance estimation method according to claim 2, characterized in that: In the time domain optimization step, the original data D RSSI The smoothing filter processing includes: according to S n =A n / M n The value of determines the smoothness of the filter window, and is used as the basis for selecting the size of the Gaussian function smoothing scale, thereby adjusting the weighting coefficient of the filter and realizing the adaptive adjustment of the filter scale parameter. The rules for adaptive adjustment of the standard deviation σ are as follows:

4. The distance estimation method according to claim 1, characterized in that: In the RSSI screening step, the step size of the partition is: (max(D LF )-min(D LF )) / 5, the interval with the largest number of RSSI occurrences is taken as the basic unit, and the screening threshold is 1 / 2 of the number of RSSI occurrences in the basic unit.

5. A distance estimation system for suppressing complex indoor RSSI fluctuations, characterized in that: include: RSSI data conditioning module, used to receive random raw received signal strength data D input sequentially over a period of time RSSI , construct a receiving window for multiple data, input one data at a time, select the median in the window, and sequentially calculate the absolute difference between two adjacent data and accumulate them to obtain the absolute difference cumulative value; the time domain optimization module is used to optimize the original data D according to the ratio of the absolute difference cumulative value to the median. RSSI Perform smoothing filtering to obtain data D gus ; Frequency domain optimization module, used to convert data D gus Transform into frequency domain data, extract the low-frequency part of RSSI in the frequency domain, and then transform into time domain data D LF ; RSSI Screening module to count D LF The number of occurrences of RSSI data in different intervals is calculated, and the data with a number of occurrences below the screening threshold is removed, and the geometric mean RD of the remaining data is calculated. RSSI That is the estimated RSSI value; A distance estimation module is used to estimate the distance based on the mathematical relationship between RSSI and the logarithmic distance path attenuation model between the paths; In the frequency domain optimization module, the data D gus Transformed into frequency domain data, extracting the low-frequency part of RSSI in the frequency domain includes: gus After Fourier transformation, the frequency domain amplitude data is converted into the data. The data group with the largest amplitude is selected as the low-frequency range. The maximum frequency of this group is set as the low-frequency cutoff point. All data below the low-frequency cutoff point are taken as the low-frequency range. The amplitude of the high-frequency part is set to 0. The method for setting the low-frequency demarcation point includes: dividing the frequency domain amplitude data into four groups with reference to the K-means clustering algorithm, and the steps of the clustering algorithm are as follows: S1: Set the number of groups k = 4, and select the amplitudes of 4 frequency points at random as the center points; S2: Calculate the numerical difference between the amplitude of all frequency points and the center point, select the center point with the smallest numerical difference as the same group, and divide the amplitude of all frequency points into k groups; S3: Calculate the average value of the data in the group, use the average value as the new center point of each group, and return to S2 until it is iterated 5 times; After the grouping is completed, the sum of the amplitudes of the four groups is compared. The group with the largest sum is the target group. The highest amplitude frequency point in this group is used as the low-frequency demarcation point, and the frequency below the demarcation point is in the low-frequency range. In the distance estimation module, the logarithmic distance path attenuation model is: RSSI=RSSI0+10μlg(d / d0)+ε, where RSSI0 is the RSSI value at unit distance d0; ε is the environmental error coefficient; μ is the path attenuation index; and d is the estimated distance.