CN113645562B - Indoor superstore intelligent fingerprint positioning method based on 5G signal - Google Patents

Indoor superstore intelligent fingerprint positioning method based on 5G signal Download PDF

Info

Publication number
CN113645562B
CN113645562B CN202110738229.5A CN202110738229A CN113645562B CN 113645562 B CN113645562 B CN 113645562B CN 202110738229 A CN202110738229 A CN 202110738229A CN 113645562 B CN113645562 B CN 113645562B
Authority
CN
China
Prior art keywords
sub
coverage area
positioning
base station
data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110738229.5A
Other languages
Chinese (zh)
Other versions
CN113645562A (en
Inventor
张晖
王炜馨
赵海涛
孙雁飞
朱洪波
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Posts and Telecommunications
Original Assignee
Nanjing University of Posts and Telecommunications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University of Posts and Telecommunications filed Critical Nanjing University of Posts and Telecommunications
Priority to CN202110738229.5A priority Critical patent/CN113645562B/en
Publication of CN113645562A publication Critical patent/CN113645562A/en
Application granted granted Critical
Publication of CN113645562B publication Critical patent/CN113645562B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02Services making use of location information
    • H04W4/023Services making use of location information using mutual or relative location information between multiple location based services [LBS] targets or of distance thresholds
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/23Clustering techniques
    • G06F18/232Non-hierarchical techniques
    • G06F18/2321Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02Services making use of location information
    • H04W4/021Services related to particular areas, e.g. point of interest [POI] services, venue services or geofences
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30Services specially adapted for particular environments, situations or purposes
    • H04W4/33Services specially adapted for particular environments, situations or purposes for indoor environments, e.g. buildings
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W64/00Locating users or terminals or network equipment for network management purposes, e.g. mobility management
    • H04W64/006Locating users or terminals or network equipment for network management purposes, e.g. mobility management with additional information processing, e.g. for direction or speed determination

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Probability & Statistics with Applications (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The application relates to an indoor superstore intelligent fingerprint positioning method based on 5G signals. The method comprises the following steps: acquiring signal data of a positioning point currently acquired by a terminal to be positioned in a target mall; performing data cleaning on the signal data of the positioning points by adopting a hybrid filtering method to obtain cleaned data; matching sub-regions for the cleaned data based on the sub-region division result of the target market in an off-line stage, and performing position calculation based on an improved adaptive weighted neighbor algorithm according to a reference point in the sub-region to obtain the positioning information of the terminal to be positioned; and feeding back the positioning information to the terminal to be positioned. The positioning accuracy of the online stage is improved, and the complexity of position matching of the online positioning stage is reduced.

Description

Indoor superstore intelligent fingerprint positioning method based on 5G signal
Technical Field
The application relates to the technical field of indoor positioning, in particular to an indoor superstore intelligent fingerprint positioning method based on 5G signals.
Background
With the continuous development of the internet, various online electronic commerce platforms and online platforms based on various living services are continuously developed and strengthened, the market share is continuously expanded, and living spaces of traditional consumption modes such as physical stores, shopping malls and the like are continuously eaten by silkworms. Various disadvantages of the traditional consumption mode are more obvious in comparison. For customers, the problems that parking and car searching are difficult in shopping in a shopping mall are solved, store information cannot be effectively touched during shopping, and the shopping experience is poor because a strange environment cannot rapidly reach a certain place. Therefore, it is necessary to introduce a more convenient positioning navigation system to the current shopping mall so as to improve the user experience of the customer.
A mall is a complex environment and in order to provide better navigation services to customers, whether shop navigation or parking lot navigation, it is first and foremost necessary to be able to determine the location of the customer. On one hand, the indoor positioning technology comprises five types of wireless signal intersection positioning, database matching positioning, dead reckoning positioning based on an inertial sensor and multi-sensor combined positioning; the database matching positioning is to perform position estimation by matching the signals acquired on site with fingerprint information in the database, and the positioning accuracy is high. On the other hand, with the gradual progress of the 5G era, large-scale deployment of 5G commercial application is started in China, small base stations are deployed indoors, signals are accurately covered within a small range due to the characteristics of small signal transmitting power and small covering radius, the 5G base stations are deployed into smart shopping malls, the advantages of low delay, large bandwidth, dense networking and the like of a 5G network are fully utilized, the positioning coverage range is improved, the time required by customer positioning can be effectively reduced, and the concept that offline large shopping malls are convenient for customers is realized.
Indoor positioning is realized by periodically sending out signals through mobile terminal equipment or a label wrist strap carried by personnel in a place covered by a network, and after receiving the signals, a base station transmits the signals to a designated positioning server. The positioning server operates a program algorithm to judge the position of the personnel. The indoor fingerprint positioning technology comprises two stages: firstly, in an off-line stage, setting a plurality of reference points for a positioning area, wherein the reference points are specific crowds in a fingerprint database in a fingerprint matching algorithm, acquiring signal characteristic values such as RSSI, LQI and CQI equivalent at each reference point, the signal characteristics are fingerprint samples of the fingerprint database in the fingerprint matching algorithm, the fingerprint samples acquired at all the reference points form a high-quality fingerprint database required by the fingerprint matching positioning algorithm, and secondly, in an on-line stage, acquiring characteristic information of a current position at an unknown position of a terminal to be positioned to form a real-time fingerprint, and then matching the fingerprint in the fingerprint database in the off-line stage by the designed fingerprint matching algorithm to acquire a final position estimation result. So far, a great deal of progress has been made in applying different signal characteristics to the field of indoor positioning, and it is also a research hotspot to improve the fingerprint matching efficiency in the online stage by sub-region division, wherein the key technology is to divide a positioning region into a plurality of sub-regions according to collected training samples by using an algorithm, and to obtain final positioning resolving coordinates by taking signals received by a client terminal as input, thereby solving the problem of complicated matching between online positioning and an offline data fingerprint database in a large indoor mall scene.
And the positioning precision of the current online stage is lower.
Disclosure of Invention
Therefore, in order to solve the technical problems, it is necessary to provide an intelligent fingerprint positioning method for an indoor large mall based on a 5G signal, which can improve the positioning accuracy at an online stage.
An indoor superstore intelligent fingerprint positioning method based on 5G signals comprises the following steps:
acquiring signal data of a positioning point currently acquired by a terminal to be positioned in a target mall;
performing data cleaning on the signal data of the positioning points by adopting a hybrid filtering method to obtain cleaned data;
matching sub-regions for the cleaned data based on the sub-region division result of the target market in an off-line stage, and performing position calculation based on an improved adaptive weighted neighbor algorithm according to a reference point in the sub-region to obtain the positioning information of the terminal to be positioned;
and feeding back the positioning information to the terminal to be positioned.
In one embodiment, the step of matching a sub-region for the cleaned data based on a result of dividing the sub-region of the target mall in an off-line stage, and performing position calculation based on an improved adaptive weighted neighbor algorithm according to a reference point in the sub-region to which the sub-region belongs to obtain the positioning information of the terminal to be positioned includes:
matching the cleaned data to a jth sub-area in a coverage area of a base station based on a sub-area division result of the target market in an off-line stage, and determining the jth sub-area as a sub-area to which the positioning point belongs;
if it is
Figure BDA0003142297610000031
b j Selecting the Manhattan distance as a weight value, and calculating the positioning coordinates of the positioning points, namely:
the manhattan distance weight of the v-th reference point in the sub-area is calculated as follows:
Figure BDA0003142297610000032
wherein, RSRP is data after cleaning, RSRP-v is signal data of the v-th reference point of the subzone to which the RSRP-v belongs, RSRP-a is signal data of the a-th reference point of the subzone to which the RSRP-a belongs, and W v A manhattan distance weight of the v-th reference point of the sub-region to which the reference point belongs;
resolving the positioning coordinates of the positioning points as follows:
Figure BDA0003142297610000033
wherein, (x, y) is the positioning coordinate of the positioning point, (x) v ,y v ) The positioning coordinate of the v reference point;
if it is
Figure BDA0003142297610000034
Carrying out weighting calculation based on the Euclidean distance and the Pearson coefficient, and calculating the Euclidean distance of each reference point in the sub-area to which the reference point belongsB in the sub-region to which the said element belongs j The coordinates of the reference points are arranged from small to large according to the Manhattan distance; minimizing the distance to d 1 As a base point, the residue b is calculated j -1 reference points to the base point in euclidean distances:
Figure BDA0003142297610000041
wherein d is v Is the Euclidean distance from the v-th reference point to the base point, (x) 1 ,y 1 ) Is a minimum distance d 1 The positioning coordinates of the reference point;
will exceed the preset distance threshold d avg Deleting the reference points to obtain the remaining O reference points;
comparing the Euclidean distances from the rest O-1 reference points to the base point, normalizing the signal data of the reference points with the same Euclidean distance by adopting a Pearson coefficient, and calculating the similarity R between the signal data of the reference points with the same Euclidean distance and the cleaned data v The formula is as follows:
Figure BDA0003142297610000042
in the formula (I), the compound is shown in the specification,
Figure BDA0003142297610000043
calculating an inner product of the cleaned data and the signal data of the v reference point after normalization processing, wherein L is the number of reference points with equal Euclidean distance in the remaining O reference points;
performing weighted calculation according to the Euclidean distances from the remaining O reference points to the base point and the similarity between the signal data of the reference points with the same Euclidean distances and the cleaned data to obtain the positioning coordinates of the positioning point, wherein the calculation formula is shown as follows;
Figure BDA0003142297610000044
and acquiring the positioning information of the terminal to be positioned according to the positioning coordinates of the positioning points and the sub-regions to which the positioning points belong.
In one embodiment, the offline stage divides the sub-areas of the target mall by:
carrying out area division on the target market by combining the deployment density of the base stations in the target market, and determining the coverage area of each base station;
analyzing according to signal data fed back by reference points received by each base station, and determining the pedestrian flow state corresponding to the coverage area of each base station;
determining a sub-area division number interval corresponding to the coverage area according to the pedestrian flow state corresponding to the coverage area of each base station;
and clustering the reference points of the coverage area by adopting a fuzzy C-means algorithm according to the sub-area division number interval corresponding to the coverage area, and judging the optimal sub-area division by combining the error square sum for analysis to obtain the sub-area division result of the coverage area.
In one embodiment, the step of analyzing according to the signal data of each base station that receives the reference point feedback to determine the pedestrian volume status corresponding to the coverage area of each base station includes:
analyzing based on a fluctuation variance formula according to signal data fed back by a reference point received by the coverage area of each base station, and determining the signal fluctuation degree of the coverage area of each base station;
the fluctuating variance equation is:
Figure BDA0003142297610000051
Figure BDA0003142297610000052
in the formula (I), the compound is shown in the specification,
Figure BDA0003142297610000053
the signal fluctuation degree of the v-th reference point of the coverage area of the base station is represented by Z, the quantity of signal characteristic values acquired by the v-th reference point in the coverage area of the base station is represented by RSRP-v i For the ith signal characteristic value of the vth reference point,
Figure BDA0003142297610000054
is the average of the signal data for the v reference point within the coverage area of that base station,
Figure BDA0003142297610000055
the signal fluctuation degree of the coverage area of the base station is H, and the reference point number of the coverage area of the base station is H;
when the signal fluctuation degree of the coverage area of the base station exceeds a preset fluctuation threshold value, the pedestrian flow state is busy;
and if the signal fluctuation degree of the coverage area of the base station is less than or equal to a preset fluctuation threshold value, the people flow state is idle.
In one embodiment, the step of determining, according to the traffic flow state corresponding to the coverage area of each base station, a sub-area division number interval corresponding to the coverage area includes:
when the pedestrian flow state of the coverage area of the base station is busy, the number of the coverage area partitions of the base station is greater than the number r of the hot spot areas and cannot exceed the number H of the reference points of the coverage area of the base station, and the sub-area partition number interval corresponding to the coverage area of the base station is [ r, H ];
when the pedestrian volume state of the coverage area of the base station is idle, performing sub-area division based on the number of shops in the coverage area of the base station, namely, the sub-area division number corresponding to the coverage area of the base station is the number of shops in the coverage area of the base station.
In one embodiment, the step of clustering the reference points of the coverage area by using a fuzzy C-means algorithm according to the sub-area division number interval corresponding to the coverage area, and determining the optimal sub-area division by combining the square sum of errors for analysis to obtain the sub-area division result of the coverage area includes:
dividing the number interval of the sub-areas corresponding to the coverage area as the class number of the fuzzy C mean value, calculating the membership degree of each reference point, initializing a membership degree matrix, and calculating an initial fingerprint clustering center through the membership degree;
updating the clustering center c according to the objective function j The objective function is:
Figure BDA0003142297610000061
in the formula, U represents the sub-area division number, H is the reference point number of the coverage area of the base station, and the membership degree is recalculated according to a new clustering center
Figure BDA0003142297610000062
c j The RSRP-v is the signal data of a v reference point in the jth sub-region, and beta is a weighting coefficient or a fuzzy control parameter;
evaluating the cluster quality by using an error cubic value formula improved based on the sum of squared errors, judging the subregion partition effect, and determining the optimal subregion partition number to obtain the optimal subregion partition;
the improved error cubic value formula based on the error square is as follows:
Figure BDA0003142297610000071
wherein D-SSE represents a cubic error value obtained by cubic computation of the sum of squared errors, c j Is the central point of the jth sub-region, RSRP-v is the signal data of the vth reference point in the jth sub-region, F j The number of reference points in the jth sub-region.
In one embodiment, the step of performing data cleaning on the signal data of the positioning point by using a hybrid filtering method to obtain cleaned data includes:
preserving the signal characteristic value in the preset interval in the signal data of the positioning point by adopting a Gaussian filtering method to obtain M 1 Calculating the mean value of the data to obtain a first estimated value rsrp 1 The preset interval is as follows:
Figure BDA0003142297610000072
wherein, the sigma is the standard deviation,
Figure BDA0003142297610000073
is the average value of the signal data of the acquired locating points.
Calculating a high-end abnormal value and a low-end abnormal value in the signal data of the positioning point by adopting a Dixon filtering method, setting the detection level alpha to be 0.05 based on a Dixon critical index, and removing an outlier to obtain M 2 Averaging the signal characteristic values to obtain a second estimation value rsrp 2
Synthesizing the first estimated value rsrp 1 And a second estimate rsrp 2 Obtaining cleaned data according to the formula;
Figure BDA0003142297610000074
wherein, RSRP is the data after cleaning.
According to the indoor superstore intelligent fingerprint positioning method based on the 5G signals, signal data of a positioning point currently acquired by a terminal to be positioned in a target mall is acquired; performing data cleaning on the signal data of the positioning points by adopting a hybrid filtering method to obtain cleaned data; matching sub-regions for the cleaned data based on the sub-region division result of the target market in an off-line stage, and performing position calculation based on an improved adaptive weighted neighbor algorithm according to a reference point in the sub-region to obtain the positioning information of the terminal to be positioned; and feeding back the positioning information to the terminal to be positioned. The positioning precision of the online stage is improved, and the complexity of position matching of the online positioning stage is reduced.
Drawings
Fig. 1 is a schematic flow chart of an intelligent fingerprint positioning method for an indoor large mall based on a 5G signal in an embodiment;
FIG. 2 is a flow chart of an offline stage subregion algorithm of the indoor mall intelligent fingerprint positioning method based on the 5G signal in one embodiment;
fig. 3 is a flow chart illustrating a position calculation in a line phase of an intelligent fingerprint positioning method for an indoor mall based on a 5G signal in an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In one embodiment, as shown in fig. 1, a method for intelligently positioning fingerprints in an indoor superstore based on 5G signals is provided, which is applied to a base station currently accessed by a terminal to be positioned in a target superstore as an execution subject, and includes the following steps:
step S220, acquiring signal data of a positioning point currently acquired by a terminal to be positioned in a target mall.
And S240, performing data cleaning on the signal data of the positioning point by adopting a hybrid filtering method to obtain the cleaned data.
And filtering the acquired signal data of the positioning points by utilizing a plurality of filtering modes. Due to the fact that the large-scale market environment has the characteristics of complexity and dynamics, personnel flow is large, 5G signals are easily interfered in the transmission process, when a terminal to be positioned is used for signal acquisition, the process of acquiring the signals each time cannot face a corresponding access point base station, when the access point base station is backed to acquire the signals, the received signals can penetrate through a human body to cause attenuation in different degrees, and the acquired signal strength has great instability; meanwhile, various metal bodies also interfere with signal values. Therefore, the selection of the signal characteristic value is very important no matter in the off-line stage or the on-line stage, and therefore, the cleaned data is obtained after the combination of the gaussian filtering and the dixon filtering.
In one embodiment, the step of performing data cleaning on the signal data of the anchor point by using a hybrid filtering method to obtain cleaned data includes:
preserving the data value in the preset interval in the signal data of the positioning point by adopting a Gaussian filtering method to obtain M 1 Calculating the mean value of the data to obtain a first estimated value rsrp 1 The preset interval is as follows:
Figure BDA0003142297610000091
wherein, the sigma is the standard deviation,
Figure BDA0003142297610000092
is the average value of the signal data of the acquired locating points.
Calculating a high-end abnormal value and a low-end abnormal value in the signal data of the positioning point by adopting a Dixon filtering method, setting the detection level alpha to be 0.05 based on a Dixon critical index, and removing an outlier to obtain M 2 Averaging the data to obtain a second estimation value rsrp 2
Synthesizing the first estimated value rsrp 1 And a second estimate rsrp 2 Obtaining cleaned data according to the formula;
Figure BDA0003142297610000093
wherein, the RSRP is the data after cleaning.
Specifically, the signal data acquired at the anchor point (the signal data includes m signal feature values) is gaussian filtered. The signal eigenvalue distribution of the 5G signal can be described by a standard normal distribution:
RSRP~N(μ,σ 2 ) I.e. by
Figure BDA0003142297610000101
Obey N (0, 1)
Standard normal distribution:
Figure BDA0003142297610000102
where μ is the mean and σ is the standard deviation.
Figure BDA0003142297610000103
Wherein, the first and the second end of the pipe are connected with each other,
Figure BDA0003142297610000104
is the average value of the signal data of the collected positioning points, m is the number of signal characteristic values in the collected signal data, RSRP i The signal characteristic value is the ith signal characteristic value in the signal data of the collected positioning point.
Selecting a signal characteristic value with the retention probability P (x) higher than 0.95, basically adapting to the change of the market environment, and inquiring a standard normal distribution table according to a density function of the standard normal distribution to obtain x =1.65:
Figure BDA0003142297610000105
where x is the value retained after Gaussian filtering is in the range x less than one standard deviation from the mean, i.e. x is
Figure BDA0003142297610000106
Where σ is the standard deviation and e is a mathematical constant with a value of about 2.72, t is a probability density function used in place of x to facilitate the determination of x.
Therefore, the temperature of the molten metal is controlled,
Figure BDA0003142297610000107
i.e. to reserve
Figure BDA0003142297610000108
The intensity value of the interval, the signal characteristic value of the positioning point acquired in a period of time obtained at the moment tends to be flat, and the residual M is calculated 1 Averaging the characteristic values of the signals to obtain a first estimated value rsrp 1 The first estimated value calculation formula is:
Figure BDA0003142297610000111
wherein rsrp 1 Fingerprint data, M, of acquired signal data obtained after Gaussian filtering 1 The number of the signal characteristic values reserved for the Gaussian filtering is that M signal characteristic values are collected at a certain point at the beginning and M is remained after the Gaussian filtering 1 A signal characteristic value.
Meanwhile, arranging m signal characteristic values of the acquired positioning points in a sequence from small to large by adopting a Dixon filtering method to obtain { RSRP 1 ,RSRP 2 ,...,RSRP m And setting the detection level alpha to 0.05, and calculating a high-end abnormal value and a low-end abnormal value in the obtained signal characteristic value sequence according to the following dixon statistical formula.
When m =3 to 7, the high-end abnormal value r 10 And low end outlier
Figure BDA0003142297610000112
The calculation is as follows:
Figure BDA0003142297610000113
wherein the high-end abnormal value r 10 The maximum signal characteristic value and the low-end abnormal value among the m =3 to 7 signal characteristic values
Figure BDA0003142297610000114
Is the smallest signal characteristic value among the m =3 to 7 signal characteristic values; RSRP 1 Is the 1 st signal characteristic value; RSRP 2 Is the 2 nd signal characteristic value.
When m =8 to 10, the high-end abnormal value r 11 And low end outliers
Figure BDA0003142297610000115
The calculation is as follows:
Figure BDA0003142297610000116
when m =11 to 13, the high-end abnormal value r 21 And low end outliers
Figure BDA0003142297610000117
The calculation is as follows:
Figure BDA0003142297610000121
when m =14 to 30, the high-end abnormal value r 22 And low end outlier
Figure BDA0003142297610000122
The calculation is as follows:
Figure BDA0003142297610000123
by looking up the critical index of dixon test, the corresponding critical value D (α, m) is obtained. If it is when
Figure BDA0003142297610000124
And r is ij >D (α, m), then RSRP in the sequence m Is determined as an outlier;
Figure BDA0003142297610000125
and is
Figure BDA0003142297610000126
Then RSRP 1 Is judged as an outlier, otherwise no outlier is judged in the sequence, wherein r ij The value of the high-end abnormal value,
Figure BDA0003142297610000127
for low-end outliers, ij is the number of high-end outliers and low-end outliers.
After the determined outliers are eliminated, the steps are repeated for the remaining signal characteristic values in the sequence until no outliers exist in the sequence, and the remaining last M is calculated 2 Calculating the average value of the characteristic values of the signals to obtain a second estimated value rsrp 2
Figure BDA0003142297610000128
Wherein M is 2 The number of the signal characteristic values reserved for the Gaussian filtering is that M signal characteristic values are collected on the positioning point at the beginning, and M signal characteristic values are remained after the Gaussian filtering 2 A signal characteristic value.
After the mixing and filtering, the final signal estimation value RSRP (i.e. the cleaned data) is:
Figure BDA0003142297610000129
and step S260, based on the division result of the sub-regions of the target market in the off-line stage, matching the sub-regions for the cleaned data, and performing position calculation based on an improved adaptive weighted neighbor algorithm according to the reference points in the sub-regions to which the sub-regions belong, so as to obtain the positioning information of the terminal to be positioned.
In the online stage, the cleaned data is divided by using the sub-region obtained by the training in the offline stage (namely, the sub-region division result of the target market based on the offline stage) to obtain the corresponding sub-region to which the terminal to be positioned currently belongs, and fingerprint matching in the sub-region is performed by using an improved self-adaptive weighted neighbor algorithm, so that the final positioning information of the terminal to be positioned and the positioning of the user holding the terminal to be positioned are realized.
Step S280, positioning information is fed back to the terminal to be positioned.
As shown in fig. 2, in an embodiment, the step of dividing the sub-area of the target mall based on the offline stage, matching the sub-area for the cleaned data, and performing position calculation based on the improved adaptive weighted neighbor algorithm according to the reference point in the sub-area to which the sub-area belongs to obtain the positioning information of the terminal to be positioned includes:
matching the cleaned data to a jth sub-area in the coverage area of the base station based on the sub-area division result of the target market in the off-line stage, and determining the jth sub-area as a sub-area to which the cleaned data belongs;
if it is
Figure BDA0003142297610000131
b j The number of reference points of the jth sub-area, H is the parameter of the coverage area of the base station
And (3) calculating the positioning coordinate of the terminal to be positioned by selecting a Manhattan distance as a weight value and taking the number of the examination points, wherein h is the number of sub-regions of a coverage area of the base station to which the examination points belong, namely:
the manhattan distance weight for the v-th reference point is calculated as follows:
Figure BDA0003142297610000132
wherein, RSRP is data after cleaning, RSRP-v is signal data of the v-th reference point of the subzone to which the RSRP-v belongs, RSRP-a is signal data of the a-th reference point of the subzone to which the RSRP-a belongs, and W v A manhattan distance weight of a vth reference point of the belonging sub-region;
the positioning coordinates of the positioning points are calculated as follows:
Figure BDA0003142297610000141
wherein, (x, y) is the positioning coordinate of the positioning point, (x) v ,y v ) Positioning coordinates for the v-th reference point;
if it is
Figure BDA0003142297610000142
Performing weighting calculation based on Euclidean distance and Pearson coefficient, calculating Euclidean distance of each reference point in the sub-region to which the reference point belongs, and performing weighting calculation on the Euclidean distance and the Pearson coefficient to obtain b j The coordinates of the reference points are arranged from small to large according to the Manhattan distance; minimizing the distance to d 1 As a base point, the residue b is calculated j 1 euclidean distance of the reference points to the base point:
Figure BDA0003142297610000143
wherein d is v Is the Euclidean distance from the v-th reference point to the base point, (x) 1 ,y 1 ) Is a distance of minimum d 1 The location coordinates of the reference point of (a);
will exceed the preset distance threshold d avg Deleting the reference points to obtain the remaining O reference points;
comparing the Euclidean distances from the rest O-1 reference points to the base point, normalizing the signal data of the reference points with the same Euclidean distance by adopting a Pearson coefficient, and calculating the similarity R between the signal data of the reference points with the same Euclidean distance and the cleaned data v The formula is as follows:
Figure BDA0003142297610000144
in the formula (I), the compound is shown in the specification,
Figure BDA0003142297610000145
calculated for the cleaned data and the signal data of the v-th reference point after normalization processingInner product, L is the number of reference points with equal Euclidean distance in the remaining O reference points;
performing weighted calculation according to the Euclidean distances from the remaining O reference points to the base point and the similarity between the signal data of the reference points with the same Euclidean distance and the cleaned data to obtain a positioning coordinate of the terminal to be positioned, wherein the calculation formula is shown as follows;
Figure BDA0003142297610000151
and acquiring the positioning information of the terminal to be positioned according to the positioning coordinates of the terminal to be positioned and the cleaned sub-region to which the data belongs.
The coverage area of the base station corresponding to the terminal to be positioned contains H reference points, the coverage area of the base station is divided into H sub-areas according to the sub-area division result of the target market through an off-line stage, and the number of the reference points in each sub-area is recorded as follows: b = { B = 1 ,b 2 ,...,b h }. And the terminal to be positioned receives the signal data for positioning, performs data cleaning based on the signal data through mixed filtering, completes the matching of sub-regions through the matching of the cleaned data and the clustering center, and matches the terminal to be positioned into the jth sub-region. The positioning information of the terminal to be positioned comprises a specific floor and a position of the terminal to be positioned, and a preset distance threshold d avg Is the mean value, i.e.:
Figure BDA0003142297610000152
as shown in fig. 3, in an embodiment, the way of dividing the sub-area of the target mall in the offline stage includes:
carrying out area division on a target market by combining the deployment density of base stations in the target market, and determining the coverage area of each base station; analyzing according to signals fed back by the reference points received by each base station, and determining the pedestrian flow state corresponding to the coverage area of each base station; determining a sub-area division number interval corresponding to a coverage area according to the pedestrian flow state corresponding to the coverage area of each base station; and clustering the reference points of the coverage area by adopting a fuzzy C-means algorithm according to the sub-area division number interval corresponding to the coverage area, and judging the optimal sub-area division by combining the error square sum for analysis to obtain the sub-area division result of the coverage area.
After the signal data fed back by the reference point is received and data cleaning is needed (the data cleaning mode is the same as the data cleaning mode of the signal data currently collected by the terminal to be positioned and is not repeated), the signal data corresponds to the position coordinates and the floors of the reference point one by one, and a high-quality off-line database is constructed to form a high-efficiency fingerprint map of the mall. In order to reduce the calculation amount of the online stage, the fingerprint map is divided into positioning sub-areas. Starting from the most basic reference point, a plurality of reference points are classified into one group, so that the coverage area of each 5G base station is divided into a plurality of sub-areas for positioning. The process of positioning the terminal to be positioned is an online stage, the sub-area division result of the target market is an offline stage, and the sub-area division result of the target market can be continuously updated according to a preset time interval in the offline stage.
In one embodiment, the step of analyzing according to the signal data of each base station receiving the feedback of the reference point to determine the pedestrian volume state corresponding to the coverage area of each base station includes:
analyzing based on a fluctuation variance formula according to signals fed back by the reference points received by the coverage area of each base station, and determining the signal fluctuation degree of the coverage area of each base station;
the fluctuation variance formula is:
Figure BDA0003142297610000161
Figure BDA0003142297610000162
in the formula (I), the compound is shown in the specification,
Figure BDA0003142297610000163
signal fluctuation degree of a v-th reference point in a coverage area of a base station, Z is the number of signal characteristic values acquired by the v-th reference point in the coverage area of the base station, RSRP-v i For the ith signal characteristic value of the vth reference point,
Figure BDA0003142297610000164
is the average of the signal data for the v reference point within the coverage area of that base station,
Figure BDA0003142297610000165
the signal fluctuation degree of the coverage area of the base station is H, and the reference point number of the coverage area of the base station is H;
when the signal fluctuation degree of the coverage area of the base station exceeds a preset threshold value, the pedestrian flow state is busy;
and if the signal fluctuation degree of the coverage area of the base station is less than or equal to a preset threshold value, the pedestrian flow state is idle.
Specifically, the method comprises the following steps: when a terminal to be positioned in coverage areas of different base stations is positioned, because the distance between the position of the terminal to be positioned and the base station corresponding to the area is equal, signal characteristic values received by the terminal to be positioned are equal, and therefore calculation of position coordinates of the terminal to be positioned is influenced. Therefore, in the off-line stage, rough division is firstly carried out according to the deployment density of the base stations in the shopping mall, and the signal path loss is estimated by combining objective environmental conditions such as floor partitions, building partitions and the like of the shopping mall, so that the deployment density of the base stations in the shopping mall is obtained. Obtaining the coverage area of each base station based on the deployment density of the base stations, finely dividing sub-areas based on floor distribution of the coverage area of the base stations, judging busy hours and idle hours in a business yard according to fluctuation conditions of signal data of reference points in the coverage area of each base station, when the traffic is small, the fluctuation range of collected signals which are not subjected to data cleaning is small, the signal intensity distribution is concentrated, when the traffic is increased, signals are obviously influenced, the fluctuation range is enlarged, the probability distribution of the signal intensity is not concentrated in a certain range, even large fluctuation occurs, analyzing is carried out based on a fluctuation variance formula, and the signal fluctuation degree of the coverage area of each base station is determined;
the fluctuating variance equation is:
Figure BDA0003142297610000171
Figure BDA0003142297610000172
in the formula (I), the compound is shown in the specification,
Figure BDA0003142297610000173
signal fluctuation degree of a v-th reference point in a coverage area of a base station, Z is the number of signal characteristic values acquired by the v-th reference point in the coverage area of the base station, RSRP-v i For the ith signal characteristic value of the v reference point,
Figure BDA0003142297610000174
is the average of the signal data of the v reference point within the coverage area of the base station,
Figure BDA0003142297610000175
the signal fluctuation degree of the coverage area of the base station is H, and the reference point number of the coverage area of the base station is H; the busy hour and the idle hour are distinguished by a preset fluctuation threshold, the preset fluctuation threshold can be set to be 5, if the signal fluctuation degree reaches the preset fluctuation threshold, the signal is considered as the busy hour, otherwise, the signal is considered as the idle hour.
In one embodiment, the step of determining the sub-area division number interval corresponding to the coverage area according to the pedestrian volume state corresponding to the coverage area of each base station includes:
when the pedestrian flow state of the coverage area of the base station is busy, the number of the coverage area partitions of the base station is greater than the number r of the hot spot areas and cannot exceed the number H of the reference points of the coverage area of the base station, and the sub-area partition number interval corresponding to the coverage area of the base station is [ r, H ]; and when the people flow state of the coverage area of the base station is idle, performing subarea division based on the number of shops in the coverage area of the base station, namely, the subarea division number corresponding to the coverage area of the base station is the number of shops in the coverage area of the base station.
When the pedestrian flow state of the coverage area of the base station is busy, the area division of the fingerprint library needs to be more dense, the number of the coverage area division of the base station is larger than the number r of the hot spot areas and cannot exceed the number H of the reference points, the probability that the optimal area division number is in an interval [ r, H ] is higher, therefore, the optimal sub-area division number H is searched in the interval, the judgment of the hot spot areas is also based on the signal fluctuation degree, and the sum of the optimal division sub-area numbers of the coverage areas of the base stations is the sub-area division number of the whole market.
And when the pedestrian flow state of the coverage area of the base station is idle, the pedestrian flow of the shopping mall is less, at the moment, sub-area division is carried out on the basis of the number of shops in the shop and the floor where the shops are located, namely the number of the divided sub-areas at the moment is set as the number of shops in the coverage area of the base station, the number of shops is determined by the area and the floor in the coverage area of the base station, and then the sub-area division of the data in the fingerprint database is completed by adopting a clustering algorithm.
In one embodiment, the step of clustering reference points of the coverage area by using a fuzzy C-means algorithm according to the sub-area division number interval corresponding to the coverage area, and determining the optimal sub-area division by combining the error sum of squares to analyze, so as to obtain the sub-area division result of the coverage area includes:
dividing the number interval of the sub-areas corresponding to the coverage area as the class number of the fuzzy C mean value, calculating the membership degree of each reference point, initializing a membership degree matrix, and calculating an initial fingerprint clustering center through the membership degree;
updating the clustering center c according to the objective function j The objective function is:
Figure BDA0003142297610000191
in the formula, U represents the sub-area division number, H is the reference point number of the coverage area of the base station, and the membership degree is recalculated according to a new clustering center
Figure BDA0003142297610000192
c j The cluster center of the jth sub-region is defined, and RSRP-v is signal data of a vth reference point in the jth sub-region;
evaluating the cluster quality by using an error cubic value formula improved based on the sum of squared errors, judging the subregion partition effect, and determining the optimal subregion partition number to obtain the optimal subregion partition;
the improved error cubic value formula based on the error square is as follows:
Figure BDA0003142297610000193
wherein D-SSE represents a cubic error value obtained by cubic calculation of the sum of squared errors, c j Is the central point of the jth sub-region, RSRP-v is the signal data of the vth reference point in the jth sub-region, F j The number of reference points in the jth sub-region.
Wherein, dividing the number interval of the corresponding sub-areas of the coverage area of each base station obtained based on the indoor large mall scene as the class number of the fuzzy C mean value, calculating the membership degree of each reference point, initializing the membership degree matrix, calculating the initial fingerprint clustering center through the membership degree, wherein each reference point has U membership degrees,
Figure BDA0003142297610000194
the membership degree of the v reference point belonging to the j category is represented, and the value range is [0, 1')]When it comes to
Figure BDA0003142297610000195
Then, the reference point belongs to category c. The degree of membership is calculated as:
Figure BDA0003142297610000196
updating the clustering center c according to the objective function j The objective function is
Figure BDA0003142297610000201
In the formula, β is a weighting coefficient or a fuzzy control parameter, and the value may be set to 2, and the lagrange multiplier method is used to solve the objective function, thereby obtaining a condition that the objective function takes a minimum value.
Figure BDA0003142297610000202
Figure BDA0003142297610000203
Wherein, c g Is the cluster center of the g-th sub-region,
updating the clustering center c according to the above formula j And according to the new cluster center c j Recalculating membership
Figure BDA0003142297610000204
Judging whether a convergence condition is satisfied, wherein the convergence condition is
Figure BDA0003142297610000205
Wherein t is the number of iteration steps; ε is a small constant that represents the error threshold. If the convergence condition is not met, updating the clustering center and the membership degree again; and if the convergence condition is reached, outputting the membership degree and the clustering center, summarizing all the reference points into each subarea according to the membership degree, and carrying out class marking.
In the application, the cluster advantages and disadvantages are evaluated by using the error Sum of Squares (SSE), and the subregion dividing effect is judged, so that the optimal subregion dividing number is determined. When the number k of the divided sub-regions is smaller than the natural clustering number (optimal clustering in a theoretical state) of the data set, the SSE is greatly reduced along with the increase of the value k; and when the k value is gradually close to the natural clustering number, the increasing degree of the polymerization degree obtained by continuously increasing k is rapidly reduced, the descending amplitude of the SSE is controlled and finally gradually becomes gentle, the relation graph of the SSE and the k is the shape of an elbow, and the k value corresponding to the elbow of the elbow is the sub-area division number of the sample data set. However, the influence of the change of the number of the sub-regions, that is, the number of the clusters, on the SSE value is not obvious, so that a cubic function with monotonically increasing positive value and sensitive to the change is introduced to improve the calculation of the error sum of squares to obtain a cubic error value D-SSE, and the specific formula is as follows:
Figure BDA0003142297610000211
wherein D-SSE represents a cubic error value obtained by cubic calculation of the error sum of squares, c j Is the cluster center of the jth sub-region, RSRP-v is the signal data of the vth reference point in the jth sub-region, F j The number of reference points in the jth sub-region. After the optimal sub-area division is completed, the fingerprint data of each reference point in the fingerprint database is { RSRP-v, l, (x) v ,y v ) Wherein RSRP-v is the signal of the v-th reference point, l is the floor on which the v-th reference point is located, (x) v ,y v ) Is the coordinate information of the v-th reference point.
According to the indoor superstore intelligent fingerprint positioning method based on the 5G signals, signal data of a positioning point currently acquired by a terminal to be positioned in a target mall is acquired; performing data cleaning on the signal data of the positioning points by adopting a hybrid filtering method to obtain cleaned data; dividing results of sub-regions of the target market based on an off-line stage, matching the sub-regions for the cleaned data, and performing position calculation based on an improved adaptive weighted neighbor algorithm according to reference points in the sub-regions to which the sub-regions belong to obtain positioning information of the terminal to be positioned; and feeding back the positioning information to the terminal to be positioned. The positioning precision of the online stage is improved, and the complexity of position matching of the online positioning stage is reduced.
Further, compared with the traditional positioning method, the indoor superstore intelligent fingerprint positioning method based on the 5G signals is intensively deployed in the intelligent mall in the form of the small base station by means of the 5G base station, the signals received by the user terminal are strong, the signal characteristic values in the signals are extracted to construct a high-quality fingerprint library, the fingerprint library is divided into a plurality of sub-regions by a method of performing region division based on the mall base station deployment and the people flow characteristics, the sub-region division enables the received signals to be matched with the sub-regions firstly, and only the positions of the received signals and the positions of reference points in the sub-regions are required to be matched as reference positions, so that the complexity of position matching in the online positioning stage is effectively reduced.
Further, based on the obtained fingerprint database constructed by the signal characteristics, the position and the sub-region, the position calculation of the positioning points is completed by using an improved self-adaptive weighting K nearest neighbor algorithm, when the number of the reference points of the sub-region is less, a Manhattan weight is adopted, otherwise, the points in the sub-region are weighted and calculated based on the Euclidean distance and the Pearson coefficient, and the accuracy of calculating the indoor positioning position is further obviously improved.
It should be understood that, although the steps in the flowchart of fig. 1 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not limited to being performed in the exact order illustrated and, unless explicitly stated herein, may be performed in other orders. Moreover, at least a portion of the steps in fig. 1 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performance of the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent application shall be subject to the appended claims.

Claims (6)

1. An indoor superstore intelligent fingerprint positioning method based on 5G signals is characterized by comprising the following steps:
acquiring signal data of a positioning point currently acquired by a terminal to be positioned in a target mall;
performing data cleaning on the signal data of the positioning points by adopting a hybrid filtering method to obtain cleaned data;
based on an off-line stage, dividing results of the sub-regions of the target market, matching the cleaned data with the sub-regions, and performing position calculation based on an improved adaptive weighted neighbor algorithm according to reference points in the sub-regions to which the cleaned data belong, so as to obtain positioning information of the terminal to be positioned;
feeding back the positioning information to the terminal to be positioned;
the step of obtaining the positioning information of the terminal to be positioned by matching sub-regions for the cleaned data based on the result of dividing the sub-regions of the target market in the off-line stage and performing position calculation based on an improved adaptive weighted neighbor algorithm according to reference points in the sub-regions to which the sub-regions belong includes the following steps:
matching the cleaned data into a jth sub-area in the coverage area of a base station based on the result of dividing the sub-areas of the target market in an off-line stage, and determining the jth sub-area as the sub-area to which the positioning point belongs;
if it is
Figure FDA0003871206150000011
b j Selecting the Manhattan distance as a weight value, and calculating the positioning coordinates of the positioning points, namely:
the manhattan distance weight of the v-th reference point in the sub-area is calculated as follows:
Figure FDA0003871206150000012
wherein, RSRP is data after cleaning, RSRP-v is signal data of the v-th reference point of the subzone to which the RSRP-v belongs, RSRP-a is signal data of the a-th reference point of the subzone to which the RSRP-a belongs, and W v A manhattan distance weight of the v-th reference point of the sub-region to which the reference point belongs;
resolving the positioning coordinates of the positioning points as follows:
Figure FDA0003871206150000021
wherein, (x, y) is the positioning coordinate of the positioning point, (x) v ,y v ) The positioning coordinate of the v reference point;
if it is
Figure FDA0003871206150000022
Performing weighting calculation based on Euclidean distance and Pearson coefficient, calculating Euclidean distance of each reference point in the sub-region to which the reference point belongs, and performing weighting calculation on the Euclidean distance and the Pearson coefficient in the sub-region to which the reference point belongs j The coordinates of the reference points are arranged from small to large according to the Manhattan distance; minimizing the distance to d 1 As a base point, the residue b is calculated j -1 reference points to the base point in euclidean distances:
Figure FDA0003871206150000023
wherein d is v Is the Euclidean distance from the v-th reference point to the base point, (x) 1 ,y 1 ) Is a minimum distance d 1 The location coordinates of the reference point of (a);
will exceed the preset distance threshold d avg Deleting the reference points to obtain the remaining O reference points;
comparing the Euclidean distances from the rest O-1 reference points to the base point, normalizing the signal data of the reference points with the same Euclidean distance by adopting a Pearson coefficient, and calculating the similarity R between the signal data of the reference points with the same Euclidean distance and the cleaned data v The formula is as follows:
Figure FDA0003871206150000024
in the formula (I), the compound is shown in the specification,
Figure FDA0003871206150000031
calculating an inner product of the cleaned data and the signal data of the v reference point after normalization processing, wherein L is the number of reference points with equal Euclidean distance in the remaining O reference points;
performing weighted calculation according to Euclidean distances from the remaining O reference points to the base point and the similarity between the signal data of the reference points with the same Euclidean distance and the cleaned data to obtain the positioning coordinates of the positioning point, wherein the calculation formula is shown as follows;
Figure FDA0003871206150000032
and acquiring the positioning information of the terminal to be positioned according to the positioning coordinates of the positioning points and the sub-regions to which the positioning points belong.
2. The method of claim 1, wherein the offline stage of dividing the sub-area of the target mall comprises:
carrying out area division on the target market by combining the deployment density of the base stations in the target market, and determining the coverage area of each base station;
analyzing according to signal data fed back by reference points received by each base station, and determining the pedestrian flow state corresponding to the coverage area of each base station;
determining a subarea division number interval corresponding to the coverage area according to the pedestrian flow state corresponding to the coverage area of each base station;
and clustering the reference points of the coverage area by adopting a fuzzy C-means algorithm according to the sub-area division number interval corresponding to the coverage area, and judging the optimal sub-area division by combining the error sum of squares to analyze so as to obtain the sub-area division result of the coverage area.
3. The method according to claim 2, wherein the step of determining the traffic status corresponding to the coverage area of each base station by analyzing the signal data of each base station received from the reference point comprises:
analyzing based on a fluctuation variance formula according to signal data fed back by receiving a reference point in the coverage area of each base station, and determining the signal fluctuation degree of the coverage area of each base station;
the fluctuating variance equation is:
Figure FDA0003871206150000041
Figure FDA0003871206150000042
in the formula (I), the compound is shown in the specification,
Figure FDA0003871206150000043
signal fluctuation degree of a v-th reference point in a coverage area of a base station, Z is the number of signal characteristic values acquired by the v-th reference point in the coverage area of the base station, RSRP-v i For the ith signal characteristic value of the v reference point,
Figure FDA0003871206150000044
is the average of the signal data of the v reference point within the coverage area of the base station,
Figure FDA0003871206150000045
the signal fluctuation degree of the coverage area of the base station is H, and the reference point number of the coverage area of the base station is H;
when the signal fluctuation degree of the coverage area of the base station exceeds a preset fluctuation threshold value, the people flow state is busy;
and if the signal fluctuation degree of the coverage area of the base station is less than or equal to a preset fluctuation threshold value, the pedestrian flow state is idle.
4. The method according to claim 3, wherein the step of determining the sub-area division number interval corresponding to the coverage area according to the traffic status corresponding to the coverage area of each base station comprises:
when the traffic state of the coverage area of the base station is busy, the number of the coverage area partitions of the base station is greater than the number r of the hot spot areas and does not exceed the number H of the reference points of the coverage area of the base station, and the sub-area partition number interval corresponding to the coverage area of the base station is [ r, H ];
and when the people flow state of the coverage area of the base station is idle, performing subarea division based on the number of shops in the coverage area of the base station, namely, the subarea division number corresponding to the coverage area of the base station is the number of shops in the coverage area of the base station.
5. The method according to claim 4, wherein the step of clustering the reference points of the coverage area by using a fuzzy C-means algorithm according to the sub-area division number interval corresponding to the coverage area, and analyzing by combining with the error sum of squares to determine the optimal sub-area division to obtain the sub-area division result of the coverage area comprises:
dividing the number interval of the sub-areas corresponding to the coverage area as the class number of the fuzzy C mean value, calculating the membership degree of each reference point, initializing a membership degree matrix, and calculating an initial fingerprint clustering center through the membership degree;
updating the clustering center c according to the objective function j The objective function is:
Figure FDA0003871206150000051
in the formula, U represents the division number of the subareas, H is the number of reference points of the coverage area of the base station, and the membership degree is recalculated according to the new clustering center
Figure FDA0003871206150000052
c j The reference signal is a clustering center of a jth sub-region, RSRP-v is signal data of a vth reference point in the jth sub-region, and beta is a weighting coefficient or a fuzzy control parameter;
evaluating the cluster quality by using an error cubic value formula improved based on the sum of squared errors, judging the subregion partition effect, and determining the optimal subregion partition number to obtain the optimal subregion partition;
the improved error cubic value formula based on the error square is as follows:
Figure FDA0003871206150000053
wherein D-SSE represents a cubic error value obtained by cubic calculation of the sum of squared errors, c j Is the central point of the jth sub-region, RSRP-v is the signal data of the vth reference point in the jth sub-region, F j The number of reference points in the jth sub-region.
6. The method of claim 1, wherein the step of performing data cleaning on the signal data of the anchor point by using a hybrid filtering method to obtain cleaned data comprises:
preserving the signal characteristic value of the positioning point in a preset interval in the signal data by adopting a Gaussian filtering method to obtain M 1 Calculating the mean value of the data to obtain a first estimated value rsrp 1 The preset interval is as follows:
Figure FDA0003871206150000061
wherein, the sigma is the standard deviation,
Figure FDA0003871206150000062
the average value is the average value of the signal data of the collected positioning points;
calculating a high-end abnormal value and a low-end abnormal value in the signal data of the positioning point by adopting a Dixon filtering method, setting a detection level alpha to be 0.05 based on a Dixon critical index, and removing outliers to obtain M 2 Averaging the characteristic values of the signals to obtain a second estimated value rsrp 2
Synthesizing the first estimated value rsrp 1 And a second estimate rsrp 2 Obtaining the cleaned data according to the formula;
Figure FDA0003871206150000063
wherein, the RSRP is the data after cleaning.
CN202110738229.5A 2021-06-30 2021-06-30 Indoor superstore intelligent fingerprint positioning method based on 5G signal Active CN113645562B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110738229.5A CN113645562B (en) 2021-06-30 2021-06-30 Indoor superstore intelligent fingerprint positioning method based on 5G signal

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110738229.5A CN113645562B (en) 2021-06-30 2021-06-30 Indoor superstore intelligent fingerprint positioning method based on 5G signal

Publications (2)

Publication Number Publication Date
CN113645562A CN113645562A (en) 2021-11-12
CN113645562B true CN113645562B (en) 2022-11-15

Family

ID=78416433

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110738229.5A Active CN113645562B (en) 2021-06-30 2021-06-30 Indoor superstore intelligent fingerprint positioning method based on 5G signal

Country Status (1)

Country Link
CN (1) CN113645562B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114826772B (en) * 2022-05-30 2024-03-08 中国联合网络通信集团有限公司 Data integrity verification system

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102573049A (en) * 2010-12-14 2012-07-11 北京智慧图科技发展有限责任公司 Method and system for indoor positioning
CN102821465A (en) * 2012-09-07 2012-12-12 哈尔滨工业大学 WLAN (Wireless Local Area Network) indoor positioning method based on subregion information entropy gain
CN106954190A (en) * 2017-05-11 2017-07-14 桂林电子科技大学 A kind of WIFI indoor orientation methods based on index mapping domain
CN110458205A (en) * 2019-07-23 2019-11-15 南京邮电大学 A kind of improved PCA clustering method based on WLAN indoor positioning
CN111726765A (en) * 2020-05-29 2020-09-29 山西大学 WIFI indoor positioning method and system for large-scale complex scene

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107547598B (en) * 2016-06-27 2020-07-24 华为技术有限公司 Positioning method, server and terminal

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102573049A (en) * 2010-12-14 2012-07-11 北京智慧图科技发展有限责任公司 Method and system for indoor positioning
CN102821465A (en) * 2012-09-07 2012-12-12 哈尔滨工业大学 WLAN (Wireless Local Area Network) indoor positioning method based on subregion information entropy gain
CN106954190A (en) * 2017-05-11 2017-07-14 桂林电子科技大学 A kind of WIFI indoor orientation methods based on index mapping domain
CN110458205A (en) * 2019-07-23 2019-11-15 南京邮电大学 A kind of improved PCA clustering method based on WLAN indoor positioning
CN111726765A (en) * 2020-05-29 2020-09-29 山西大学 WIFI indoor positioning method and system for large-scale complex scene

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
一种RSS和CSI融合的二阶段室内定位方法;党小超等;《传感技术学报》;20180831(第08期);全文 *
一种基于多维标度和区域细化的无线室内定位方法;肖亚龙等;《计算机学报》;20160404(第08期);全文 *
基于近邻传播算法的动态自适应室内指纹定位算法;张力仁等;《计算机应用研究》;20161012(第10期);全文 *

Also Published As

Publication number Publication date
CN113645562A (en) 2021-11-12

Similar Documents

Publication Publication Date Title
Song et al. A novel convolutional neural network based indoor localization framework with WiFi fingerprinting
CN110012428B (en) Indoor positioning method based on WiFi
Song et al. Cnnloc: Deep-learning based indoor localization with wifi fingerprinting
CN102111872B (en) Location method and device based on fingerprint location technology
CN112135248B (en) WIFI fingerprint positioning method based on K-means optimal estimation
CN106851571B (en) Decision tree-based rapid KNN indoor WiFi positioning method
CN108882172B (en) Indoor moving trajectory data prediction method based on HMM model
CN110049549B (en) WiFi fingerprint-based multi-fusion indoor positioning method and system
CN107087256A (en) A kind of fingerprint cluster method and device based on WiFi indoor positionings
CN107182036A (en) The adaptive location fingerprint positioning method merged based on multidimensional characteristic
CN112887902A (en) Indoor positioning method of WiFi fingerprint based on Gaussian clustering and hybrid measurement
CN109348416B (en) Fingerprint indoor positioning method based on binary k-means
CN108566675B (en) WiFi indoor positioning method based on multiple access point selection
CN113645562B (en) Indoor superstore intelligent fingerprint positioning method based on 5G signal
CN107290714B (en) Positioning method based on multi-identification fingerprint positioning
CN109121081B (en) Indoor positioning method based on position candidate set and EM algorithm
CN108919182B (en) Target positioning method based on support set and expectation maximization in WIFI environment
CN108445443B (en) KNN-based fingerprint point clustering method
CN110933631B (en) Indoor positioning method based on WIFI position fingerprint
CN113518308B (en) Optimal AP screening method in indoor positioning
CN115226027A (en) WiFi indoor fingerprint positioning method and device
CN108462940A (en) A kind of indoor orientation method based on Wi-Fi hotspot
CN111918211B (en) Indoor fingerprint positioning method based on KD tree and improved weighted KNN
CN113099386A (en) Multi-floor indoor position identification method and application thereof in museum navigation
Yuan et al. Research on intelligent indoor location method based on Wi-Fi fingerprint

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant