CN109379711B - positioning method - Google Patents

positioning method Download PDF

Info

Publication number
CN109379711B
CN109379711B CN201811559530.4A CN201811559530A CN109379711B CN 109379711 B CN109379711 B CN 109379711B CN 201811559530 A CN201811559530 A CN 201811559530A CN 109379711 B CN109379711 B CN 109379711B
Authority
CN
China
Prior art keywords
region
received signal
rss
data
received
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
CN201811559530.4A
Other languages
Chinese (zh)
Other versions
CN109379711A (en
Inventor
尹峰
陆彦辉
崔曙光
罗智泉
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.)
Shenzhen Big Data Research Institute
Chinese University of Hong Kong Shenzhen
Original Assignee
Shenzhen Big Data Research Institute
Chinese University of Hong Kong Shenzhen
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 Shenzhen Big Data Research Institute, Chinese University of Hong Kong Shenzhen filed Critical Shenzhen Big Data Research Institute
Priority to CN201811559530.4A priority Critical patent/CN109379711B/en
Publication of CN109379711A publication Critical patent/CN109379711A/en
Application granted granted Critical
Publication of CN109379711B publication Critical patent/CN109379711B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/029—Location-based management or tracking services
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/021—Services related to particular areas, e.g. point of interest [POI] services, venue services or geofences
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/33—Services specially adapted for particular environments, situations or purposes for indoor environments, e.g. buildings
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/80—Services using short range communication, e.g. near-field communication [NFC], radio-frequency identification [RFID] or low energy communication
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
    • H04W64/003—Locating users or terminals or network equipment for network management purposes, e.g. mobility management locating network equipment

Abstract

The invention discloses a positioning method, which comprises the following steps: determining a region to be positioned, and acquiring signal intensity data of signal transmitting equipment received by a receiving point located in the region to be positioned to obtain a received signal intensity data set of the region to be positioned, wherein the receiving point comprises a sampling point and a reference point; acquiring signal intensity data of signal transmitting equipment received by an object to be positioned to obtain received signal intensity data of the object to be positioned; and comparing the received signal strength data of the object to be positioned with the received signal strength data set of the area to be positioned to obtain the position information of the object to be positioned. The technical scheme can realize high-precision indoor positioning, and the cellular mobile network can realize positioning based on RSS, so the scheme is suitable for all positioning systems which take RSS as a position-related measured value.

Description

Positioning method
Technical Field
The invention relates to the field of signal processing, in particular to a positioning method.
Background
With the introduction of the concept of internet plus and the maturity of the technology of internet of things, the location service plays an increasingly important role in various aspects such as logistics transportation, production, emergency treatment, positioning, cruising and query. Many location services are highly dependent on the development of location technology, particularly indoor location technology. Early technologies applied to indoor positioning include Bluetooth (Bluetooth), Infrared (Infrared), and Ultra Wide Band (UWB). However, the indoor high-precision positioning using such technologies requires a large amount of hardware support, and the signal transmission process is susceptible to interference from other factors. In recent years, the wide deployment of Wi-Fi access points enables Wi-Fi signals to cover almost any position of each building, and in addition, the Wi-Fi signals are slightly interfered by the environment, so that the indoor positioning technology based on Wi-Fi is gradually a hot spot of research.
indoor positioning based on Wi-Fi is mainly divided into two categories, namely positioning algorithm based on a ranging model and positioning algorithm based on Received Signal Strength (RSS). The positioning algorithm based on the ranging model needs to perform channel estimation on the indoor channel environment in advance to establish a proper channel model. Due to the characteristic that the indoor environment is complex and changeable, after a proper channel model is determined, the indoor positioning method based on the ranging model cannot be applied to a new environment, and in a complex environment, the indoor positioning algorithm based on the ranging model has great limitation. The positioning algorithm based on RSS has the outstanding advantages of being suitable for complex indoor environments, low in calculation complexity, high in running speed and easy to achieve.
the Wi-Fi based RSS positioning process includes two phases of off-line training and on-line positioning. The main work of the off-line training stage is to construct an RSS fingerprint database, wherein the database comprises the position coordinates of a reference point and RSS information of surrounding Wi-Fi Access Points (APs) received at the reference point, and the RSS information is called RSS fingerprints; the main work of the on-line positioning stage is as follows: and measuring the RSS fingerprint information of the to-be-positioned point, and comparing the RSS fingerprint information with the RSS fingerprint information in the database by using a specific fingerprint matching algorithm to estimate the position coordinate of the to-be-positioned point.
in the indoor positioning algorithm based on Wi-Fi RSS fingerprints, whether the RSS fingerprint database is successfully constructed or not greatly influences the positioning precision and the positioning efficiency of the positioning algorithm. Currently, the main methods for constructing RSS fingerprint databases are: reference points are set in a positioning area according to a certain mode, and then RSS fingerprint information of all the reference points is collected. The method can set the number and the positions of the reference points, and the acquired information is comprehensive, so the method is mainly used in smaller indoor environments such as offices, families, movie theaters and the like, and is more suitable for occasions needing rapid deployment. However, for large public areas such as hospitals, shopping malls, parks, and the like, there are many WiFi hotspots, and the number and the positions dynamically change with time, setting up reference points according to a certain rule and collecting RSS information is a time-consuming, labor-consuming and high-cost work, and at this time, in order to save time and cost, a fingerprint map is usually obtained by measuring RSS fingerprints by walking back and forth in a positioning area according to a set route. However, in this way, only RSS fingerprints for a small number of routes are typically available, and the quality of the fingerprint database is often poor.
Disclosure of Invention
In order to improve the quality of a fingerprint database and improve positioning accuracy and positioning efficiency, the invention provides a positioning method, wherein a received signal intensity data set of an area to be positioned, namely an RSS probability fingerprint map/fingerprint database, is constructed based on a distributed recursion Gaussian process, and high-accuracy indoor positioning can be realized based on the database.
the positioning method comprises the following steps:
Determining a region to be positioned, and acquiring signal intensity data of signal transmitting equipment received by a receiving point located in the region to be positioned to obtain a received signal intensity data set of the region to be positioned, wherein the receiving point comprises a sampling point and a reference point;
acquiring signal intensity data of signal transmitting equipment received by an object to be positioned to obtain received signal intensity data of the object to be positioned;
Comparing the received signal strength data of the object to be positioned with the received signal strength data set of the area to be positioned to obtain the position information of the object to be positioned;
the determining a region to be positioned and acquiring signal intensity data of signal transmitting equipment received by a receiving point located in the region to be positioned to obtain a received signal intensity data set of the region to be positioned includes:
Determining a region to be positioned, and dividing the region to be positioned into two or more sub-regions to be positioned;
acquiring signal intensity data of signal transmitting equipment received by a receiving point positioned in the sub-region to be positioned to obtain a received signal intensity data set of the sub-region to be positioned;
Performing data fusion on the received signal intensity data set of the sub-region to be positioned to obtain the received signal intensity data set of the sub-region to be positioned;
wherein the received signal strength data for the reference point is calculated using the following equation:
Wherein it is assumed that the area to be located has L APs and that this area is divided into J sub-areas, p (RSS)l,j(p)|p,Dl,j) RSS data RSS from the ith signal transmitting device received at reference point p representing the jth sub-regionl,j(p) a posterior probability density function obtained from the sampled data of the sample points based on a recursive Gaussian process, Dl,jA training data set composed of sampling data representing sampling points of the jth sub-region, the training data in the training data set being obtained sequentially in time order,represents p (RSS)l,j(p)|p,Dl,j) Obey mean value of mul,j(p) variance ofA gaussian distribution of the intensity of the light beam,as RSSl,j(p) variance of the noise term.
In an embodiment of the present invention, the sub-regions to be located overlap or do not overlap each other.
In an embodiment of the present invention, the obtaining signal intensity data of the signal transmitting device received by the receiving point located in the sub-region to be positioned to obtain a received signal intensity data set of the sub-region to be positioned includes:
Acquiring geographic position information of a sampling point, a reference point and signal transmitting equipment in the sub-region to be positioned;
Acquiring signal intensity data of the signal transmitting equipment received at the sampling point;
And calculating to obtain reference point received signal intensity data according to the sampling point received signal intensity data to obtain a received signal intensity data set of the sub-region to be positioned.
In an embodiment of the present invention, the received signal strength data is received signal strength probability data.
In an embodiment of the present invention, the signal intensity data of the signal transmitting device received by the receiving point in the area to be located is represented as:
wherein, p (RSS)l(p)|p,Dl,1,Dl,2,,…,Dl,J,) Representing the probability density function of the received signal strength received by the reception point p from the l-th signal transmitting device, subject to the mean valuevariance ofGaussian distribution ofRSSl(p) RSS, D, for the l-th AP received at reception point pl,1,Dl,2,,…,Dl,J,And representing the received signal strength data of the ith AP measured by the sampling point in the jth sub-region, namely the training data of the jth sub-region corresponding to the ith AP, wherein J is the total number of the regions.
In one embodiment of the present invention, the mean valueCan be calculated using the following equation:
Variance (variance)Can be calculated using the following equation:
Wherein the content of the first and second substances,the uncertainty is represented by a representation of the time,As RSSlVariance of noise term in (p), Kl,j(p) and μl,jAnd (p) respectively calculating the variance and the mean of the ith AP of the jth sub-region at the time t by using a recursive Gaussian process.
In an embodiment of the present invention, the comparing the received signal strength data of the object to be positioned with the received signal strength data set of the area to be positioned to obtain the position information of the object to be positioned includes:
Calculating the distance between the received signal strength data of the object to be positioned and the received signal strength data in the received signal strength data set of the area to be positioned;
and determining the position information of the receiving point corresponding to the received signal strength data meeting the preset distance condition as the position information of the object to be positioned.
In an embodiment of the present invention, the distance is a euclidean distance.
As described above, the technical solution can achieve high-precision indoor positioning, and actually, since a cellular mobile network can also achieve positioning based on RSS, the positioning method provided by the present invention is applicable to all positioning systems using RSS as a position-related measurement value.
Drawings
FIG. 1 is a schematic view of a region to be located according to an embodiment of the present invention;
fig. 2 is a flowchart of a positioning method according to an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to specific embodiments and the accompanying drawings.
In order to make the technical solution of the present invention easier to understand, first, a brief description is made of the RSS positioning principle by which the technical solution of the present invention is implemented.
First, system model and algorithm principle
assuming that an area to be positioned where an object to be positioned (or an object to be positioned, a positioning target) is located is as shown in fig. 1, in order to position the target in the area to be positioned, an interesting grid may be defined in advance, a plurality of access points AP (such as WiFi hotspots) exist in the grid, and an intersection point of the grid is a reference point, wherein a density degree of the grid may be determined according to a requirement of an actual application, a positioning accuracy, and a system storage capability, and the present invention is not particularly limited. The RSS fingerprint FP (p) of any reference point in the grid is recorded as the set of RSS of surrounding APs received by the reference point, as shown in formula (1).
FP(p)＝{RSS1(p),RSS2(p),…,RSSL(p)} (1)
Where p is the position coordinate of the reference point, usually expressed in three-dimensional coordinates, and thus a vector; RSS (really simple syndication) data base)l(p) denotes the RSS of the L-th AP received at the reference point p, where p is 1 … L. In an embodiment of the present invention, the fingerprint FP (p) of the reference point p only contains RSS of L APs closest to (or strongest in strength), and actually, the number of APs in the area to be located may be much larger than L.
if the RSS fingerprints of all reference points are known, the target location can be achieved using a fingerprint matching algorithm (e.g., KNN, WKNN, MD5-KNN, etc.). Here, the database composed of RSS fingerprints of all reference points may also be referred to as RSS map, written as formula (2).
FPmap＝{FP(p1),FP(p2),…,FP(pn),…,FP(pN)} (2)
assuming that there are N reference points in the region to be located, in formula (2), pnposition coordinates representing the nth reference point, FP (p)n) Fingerprint data representing the nth reference point.
Obviously, when N is very large, it is not practical to directly acquire fingerprints of all reference points by measurement, and therefore, in an embodiment of the present invention, M reference points, that is, sampling points, are selected from all reference points to sample, and the corresponding RSS is used as training data, and then the probability distribution of RSS of all other reference points in the grid is calculated according to the training data, that is, a part of cross points in the grid is taken as sampling points, RSS signals of surrounding APs received at the sampling points are acquired, and RSS signals of surrounding APs received at other reference points are calculated according to the acquired data. In this embodiment, since RSS data is replaced by RSS probability distributions, the set of RSS probability distributions for surrounding APs received at reference point p, also referred to as probability fingerprints, can be written as:
FPPr(p)＝{p(RSS1(p)),p(RSS2(p)),…,p(RSSL(p))} (3)
Then the RSS probability fingerprints of all reference points constitute an RSS probability fingerprint database, written as:
FPPr_map＝{FPPr(p1),FPPr(p2),…,FPPr(pn),…,FPPr(pN)} (4)
the following describes how to calculate probability fingerprints, construct RSS probability fingerprint maps and make them suitable for the environment where training data is updated in real time and for the distributed system to perform positioning calculation.
Second, RSS probability fingerprint map is constructed based on standard Gaussian process
Generally, the RSS model of Wi-Fi signals in an indoor wireless environment can be represented by equation (5):
RSSl(p)＝PLl(p)+el(p)+nl，l＝1,…,L (5)
wherein the RSSl(p) denotes the RSS of the L-th AP received at reference point p, p ═ 1 … L; PLl(p) is the deterministic path loss function of the ith AP to the reference point p. For WiFi indoor signals, PLl(p) following an empirical path loss model:
In the formula (6), Alis at a reference distance d0at the measured transmission power value, Blis the path loss exponent, dl(p) as reference point p to coordinate pap,lEuclidean distance of the ith AP of (1):For other wireless environments, the path loss model may also be constructed from other more complex models.
N in formula (5)lIs a noise item independent of position, and is expressed by independent and identically distributed Gaussian white noise, the mean value of the noise item is zero, and the variance of the noise item isel(p) denotes noise due to shadowing effects, which are the main cause of slow fading and are usually location dependent. Thus, the noise term el(p) can be represented by a gaussian process with a mean value of zero:
Here, the first and second liquid crystal display panels are,Representing a Gaussian process, a variance function (which may also be called a kernel function) kl(p, p') can be used in any suitable form, such assay square exponential kernel functions or materrn kernel functions. In the following description of the invention, the square exponential kernel is used:
Wherein the content of the first and second substances,representing uncertainty of the model, CLlis a characteristic length scale used to express the spatial correlation between two positions, i.e. the correlation between an arbitrary position p' and another position p.
in general, the parameters in equations (6) to (8)Is unknown, so the RSS cannot be directly calculated from equation (5).
to compute a fingerprint or probabilistic fingerprint of an arbitrary reference point p on the grid, let us assume that the RSS of the ith AP has been measured (or somehow computed/estimated), denoted D, at M reference points (called sample points, sampled from N reference points) whose locations are knownl：
In this embodiment, equation (9) may be used as the training data, where plmrepresenting the position coordinates, RSS, of the m-th sample point corresponding to the l-th APl(pl1) Indicating the RSS of the i-th AP received at the m-th sample point. It should be noted that M sampling points corresponding to different APs may be different or the same, and for convenience of understanding, p is used uniformlylmTo indicate.
obviously, from the training data in the equation (9), the unknown parameter θ in the equations (6) to (8) can be estimatedlThen, the RSS of the ith AP received at any reference point can be calculated by using equation (5), but the calculation has a large error. Is connected withNext, we use another approach, namely to assume that the parameter θ is already knownlThen, a posterior probability density function thereof is calculated based on a standard Gaussian process, and then an RSS probability fingerprint database is constructed according to the probability density function thereof.
the gaussian process is a non-parametric learning model based on kernel functions. Compared with other random processes, the method has the advantages that a part of variables are randomly extracted from random variables of the Gaussian process, the process formed by the obtained variables is still Gaussian, and the joint distribution of the variables still conforms to multidimensional Gaussian distribution; in a gaussian process, each point in the input space is associated with a random variable that follows a gaussian distribution, and the joint probability of any finite number of data fusions of these random variables also follows a gaussian distribution. The gaussian process is also characterized in that it is uniquely defined by a mean function and a covariance, and a gaussian distribution corresponds to a kernel function, i.e., a covariance function. Therefore, in the model based on the gaussian process, it is generally assumed that the training data sample obeys normal distribution, and a specific gaussian process model is obtained only by obtaining a mean function and a covariance function matrix of the model. The covariance function is usually a kernel function, and the model type and performance of the gaussian process can be determined by selecting the kernel function type.
Based on the above characteristics of the gaussian process, it is apparent that the RSS observed in equation (9) obeys the following distribution:
All symbols contained in formula (10) are defined as follows:
IMIs a unit square matrix with dimension M multiplied by M.
Thus, the posterior probability density function of the RSS measurements for any reference point p on the grid can be calculated as follows:
in the formula (16), the compound represented by the formula,
According to the above formula, RSS posterior probability distributions of all L APs at the reference point p can be calculated, so as to obtain corresponding posterior probability fingerprints:
FPPr_post(p)＝{p(RSS1(p)|p,Dl),p(RSS2(p)|p,Dl),…,p(RSSL(p)|p,Dl)}
(19)
by utilizing the posterior probability fingerprints, a probability fingerprint database can be constructed to replace an RSS database, and accurate positioning can be realized by further utilizing an MD5-KNN algorithm. For the standard gaussian process, the probability density function is completely determined by the mean and covariance, so that the probability density function in the probability fingerprint can be replaced with the corresponding mean and covariance when actually processed.
Because in this method the method is carried out in batchesAll the training data are processed by the formula, so the training parameter corresponding to the method has the complexity of operation timeThe complexity of the computation time in computing the posterior probability isIn addition, the required storage space also equalsIs in direct proportion.
third, RSS probability fingerprint map is determined based on distributed recursion Gaussian process
In the second part, it can be seen that, when the probability fingerprints are calculated based on the standard gaussian process, the training data are processed in a batch mode, when the data are more and increase along with time, the required computation amount and storage amount are very large, the training process consumes a long time, and a large delay is generated for some real-time applications, so that the trial effect of the real-time applications is influenced.
In an embodiment of the invention, a distributed recursion Gaussian process is used to replace a standard Gaussian process to calculate the probability fingerprint, the use of the distributed recursion Gaussian process can effectively reduce the operation complexity, and meanwhile, the distributed recursion algorithm can also mine the time-space correlation in the data, thereby being more suitable for the processing of space-time data. In the embodiment, data are sequentially processed according to a time sequence based on a recursive Gaussian process, and then are combined with distributed data processing to realize data fusion, so that the complexity of operation time is greatly reduced, and the storage space is saved. The use of the distributed recursive gaussian process is described in detail in two steps below.
1. Recursive gaussian process
In an actual working environment, training data are usually obtained sequentially in a time sequence, and a recursive gaussian process is a gaussian process for sequentially processing the training data in the time sequence.
Suppose that time t is samplingPoint pl,tthe new set of data collected is pl,t,RSSl(pl,t) From the introduction in the second section, it can be assumed that the RSS posterior probability density function from the i-th AP received at reference point p before time t is known and is:
in the formula (20), mul,t-1(p) and Kl,t-1(p) is known, and
New data p is measuredl,t,RSSl(pl,t) After, the RSS probability density function can be updated as:
as can be seen from the above, the recursive Gaussian process is based on μl,t-1(p) and Kl,t-1(p) calculating μl,t(p) and Kl,tProcess of (p), μl,t(p) and Kl,t(p) can be represented as:
Kl,t(p)＝Kl,t-1(p)-Kl,t-1(p)gl,tJTl,t (25)
the calculation method of each parameter in the equations (24) and (25) is as follows:
The recursive algorithm reduces the temporal complexity of the overall algorithm to that of a standard gaussian process
2. Distributed recursive Gaussian process
When the RSS probability fingerprint database is constructed based on the standard Gaussian process and the recursive Gaussian process, uniform kernel functions and hyper-parameters are adopted in all the areas, and the problem is not great to some simple indoor environments (such as open indoor environments). However, for a complex indoor environment, because the interference suffered in the signal transmission process is different, different local areas may have different gaussian process parameters, i.e. different kernel functions and hyper-parameters, and therefore, it is more appropriate to divide the whole area into small areas for processing. In addition, if the data volume of one region is large, the requirement on the operation complexity of the Gaussian process is high, relatively speaking, the calculation of the posterior probability mean value and the variance of each grid in the local region is more accurate, and the large region is divided, so that considerable calculation and communication resources can be saved.
therefore, in an embodiment of the invention, an RSS probability fingerprint model is obtained based on a distributed recursive Gaussian process, and a positioning method is carried out, so that the positioning problem is better solved. As shown in fig. 2, the method comprises the steps of:
Step S1, determining a region to be positioned, and acquiring signal intensity data of a signal transmitting device received by a receiving point located in the region to be positioned to obtain a received signal intensity data set of the region to be positioned, wherein the receiving point comprises a sampling point and a reference point;
In an embodiment of the present invention, the received signal strength data is received signal strength probability data, and the received signal strength data set of the region to be located is the above-mentioned probability fingerprint database.
in an embodiment of the present invention, the step S1 includes the following steps:
Step S11, determining a region to be positioned, and dividing the region to be positioned into two or more sub-regions to be positioned;
by utilizing a distributed recursive algorithm, the whole area needs to be divided into a plurality of local areas, the local areas are divided in various ways, flexible division can be performed according to specific application requirements, and the method is not particularly limited.
In actual operation, terminals such as mobile terminals and the like can be used for collecting data in each local area, and if a plurality of mobile terminals are available, the plurality of mobile terminals can be used for collecting data simultaneously. After the data acquisition is finished, one implementation scheme is that the RSS probability fingerprints of all reference points in the area are calculated by the mobile terminal by using local area data based on a recursive Gaussian process, an RSS map of the area is constructed, and then the map is sent to a data center corresponding to the mobile terminal for data fusion, so that the map construction of the whole area is completed. Another embodiment is that after the mobile terminal finishes data acquisition, the acquired local area data is immediately sent to a data center corresponding to the mobile terminal, the RSS probability fingerprints of all reference points in the area are calculated in the data center based on a recursive Gaussian process, an RSS map based on the local area data is constructed, and finally, the obtained RSS maps of all the local area data are fused in the data center, so that a final RSS map is obtained. Because the map contains data information from each area, the map can be used as a basis for subsequent positioning.
In an embodiment of the present invention, assuming that the region to be located can be divided into J adjacent regions, the training data set D in formula (9)lCan be divided into smaller data sets Dl,j，j＝1,2,…,J。
step S12, acquiring signal intensity data of the signal transmitting equipment received by the receiving point in the sub-area to be positioned, and acquiring a received signal intensity data set of the sub-area to be positioned;
From the above, the RSS probability fingerprint can be calculated using a recursive gaussian process, and according to equation (16) or (22), the RSS probability density function received from the ith AP by the reference point p of the jth sub-region can be expressed as:
Wherein it is assumed that the area to be located has L APs and that this area is divided into J sub-areas, p (RSS)l,j(p)|p,Dl,j) RSS data RSS from the ith signal transmitting device received at reference point p representing the jth sub-regionl,j(p) a posterior probability density function obtained from the sampled data of the sample points based on a recursive Gaussian process, Dl,jA training data set composed of sampling data representing sampling points of the jth sub-region, the training data in the training data set being obtained sequentially in time order,represents p (RSS)l,j(p)|p,Dl,j) Obey mean value of mul,j(p) variance ofA gaussian distribution of the intensity of the light beam,As RSSl,j(p) variance of the noise term.
In particular for each sub-region, it is known, considering a recursive gaussian process: p (RSS)l,j(p)|p,Dl,j) Corresponding to p (RSS)l,j(p)|p,Dl,j,t)，μl,j(p) corresponds to μl,j,t(p)，Kl,j(p) corresponds to Kl,j,t(p)，Dl,jIs equivalent to Dl,j,tand D isl,j,tIs the training data set of the ith AP measured by the sub-region j at time t, please refer to the following formula corresponding to formula (23):
for each sub-region j, its p (RSS)l,j(p)|p,Dl,j,t)，μl,j,t(p)，Kl,j,tThe calculation method of (p) can be referred to p (RSS) in equations (22), (24) and (25), respectivelyl,j(p)|p,Dl,1:t)，μl,t(p)，Kl,t(p) calculation method:
Kl,j,t(p)＝Kl,j,t-1(p)-Kl,j,t-1(p)gl,tJTl,tcorresponding formula (25)
Wherein, { pl,t,RSSl,j(pl,t) At time t is sampled point pl,ta set of new data collected; mu.sl,j,t-1(p) and Kl,j,t-1(p) is the parameter of the last time, because of the recursive process, the value of 0 th time can be obtained by initialization;
Representing uncertainty of the model, CLlIs a characteristic length scale for expressing the spatial correlation between two positions;
Wherein A islIs at a reference distance d0measured transmission power value, Blis the path loss exponent, dl(p) as reference point p to coordinate pap,lEuclidean distance of the ith AP of (1):
More specifically, in combination with the above, the step S12 may include the following steps:
Step S121, acquiring geographic position information of a sampling point, a reference point and signal transmitting equipment in the sub-area to be positioned, wherein the signal transmitting equipment can be an AP (access point) such as a Wifi hotspot;
step S122, acquiring signal intensity data of the signal transmitting equipment received at the sampling point;
Step S123, calculating to obtain reference point received signal intensity data according to the sampling point received signal intensity data, namely the training data, and further obtaining a received signal intensity data set of the sub-region to be positioned.
And step S13, carrying out data fusion on the received signal intensity data set of the sub-region to be positioned to obtain the received signal intensity data set of the sub-region to be positioned.
After obtaining the posterior probability equations of the local regions, the posterior probability equations of the local regions can be fused by using a Bayesian Committee Mechanism (BCM), and the global posterior probability mean and variance of each grid after fusion can be represented by equations (30) and (31), respectively:
wherein, the upper labelf represents the quantity obtained after data fusion, betaj(p) as a function of reference point position:
Then, after data fusion, the RSS probability density function received by the reference point p from the ith AP can be expressed as:
Step S2, acquiring signal intensity data of a signal transmitting device received by an object to be positioned to obtain received signal intensity data of the object to be positioned;
step S3, comparing the received signal strength data of the object to be positioned with the received signal strength data set of the area to be positioned, to obtain the position information of the object to be positioned.
In an embodiment of the present invention, the step S3 includes:
Step S31, calculating a distance between the received signal strength data of the object to be positioned and the received signal strength data of the received signal strength data set of the area to be positioned;
step S32, determining the position information of the receiving point corresponding to the received signal strength data satisfying the preset distance condition as the position information of the object to be positioned.
the distance can be equal to the Euclidean distance, and the preset distance condition can be the minimum distance or other distance constraint conditions.
it can be seen from the above that the computation complexity of the fingerprint database obtained based on the distributed recursive gaussian process is greatly reduced, and meanwhile, the time required by the RSS measurement process of each region can be greatly shortened due to the mutual independence of the J regions. As described above, the technical solution can achieve high-precision indoor positioning, and actually, since a cellular mobile network can also achieve positioning based on RSS, the positioning method provided by the present invention is applicable to all positioning systems using RSS as a position-related measurement value.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the present invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (8)

1. A method of positioning, the method comprising the steps of:
determining a region to be positioned, and acquiring signal intensity data of signal transmitting equipment received by a receiving point located in the region to be positioned to obtain a received signal intensity data set of the region to be positioned, wherein the receiving point comprises a sampling point and a reference point;
acquiring signal intensity data of signal transmitting equipment received by an object to be positioned to obtain received signal intensity data of the object to be positioned;
Comparing the received signal strength data of the object to be positioned with the received signal strength data set of the area to be positioned to obtain the position information of the object to be positioned;
The determining a region to be positioned and acquiring signal intensity data of signal transmitting equipment received by a receiving point located in the region to be positioned to obtain a received signal intensity data set of the region to be positioned includes:
determining a region to be positioned, and dividing the region to be positioned into two or more sub-regions to be positioned;
Acquiring signal intensity data of signal transmitting equipment received by a receiving point positioned in the sub-region to be positioned to obtain a received signal intensity data set of the sub-region to be positioned;
Performing data fusion on the received signal intensity data set of the sub-region to be positioned to obtain the received signal intensity data set of the sub-region to be positioned;
Wherein the received signal strength data for the reference point is calculated using the following equation:
Wherein it is assumed that the area to be located has L APs and that this area is divided into J sub-areas, p (RSS)l，j(p)|p，Dl，j) RSS data RSS from the ith signal transmitting device received at reference point p representing the jth sub-regionl，j(p) a posterior probability density function obtained from the sampled data of the sample points based on a recursive Gaussian process, Dl，jA training data set composed of sampling data representing sampling points of the jth sub-region, the training data in the training data set being obtained sequentially in time order,Represents p (RSS)l，j(p)|p，Dl，j) Obey mean value of mul，j(p) variance ofA gaussian distribution of the intensity of the light beam,As RSSl，j(p) variance of the noise term.
2. The method according to claim 1, characterized in that the subregions to be located overlap or do not overlap each other.
3. The method of claim 1, wherein the obtaining signal strength data of signal transmitting equipment received by a receiving point located in the sub-region to be positioned to obtain a received signal strength data set of the sub-region to be positioned comprises:
Acquiring geographic position information of a sampling point, a reference point and signal transmitting equipment in the sub-region to be positioned;
Acquiring signal intensity data of the signal transmitting equipment received at the sampling point;
and calculating to obtain reference point received signal intensity data according to the sampling point received signal intensity data to obtain a received signal intensity data set of the sub-region to be positioned.
4. a method according to any of claims 1 to 3, wherein the received signal strength data is received signal strength probability data.
5. The method of claim 4, wherein the signal strength data of the signal transmitting device received by the receiving point in the area to be located is represented as:
Wherein, p (RSS)l(p)|p，Dl，1，Dl，2，，...，Dl，J，) Representing the probability density function of the received signal strength from the ith signal transmitting device received by the reception point p after data fusion, with the mean valuevariance ofGaussian distribution ofRSSl(p) RSS, D, for the l-th AP received at reception point pl，1，Dl，2，，...，Dl，J，and the received signal strength data of the ith AP measured by the sampling point in the jth sub-area is represented, namely the training data of the ith AP corresponding to the jth sub-area.
6. the method of claim 5, wherein the mean value isCan be calculated using the following equation:
variance (variance)Can be calculated using the following equation:
Wherein the content of the first and second substances, The uncertainty is represented by a representation of the time,as RSSlvariance of noise term in (p), Kl，j(p) and μl，jAnd (p) respectively calculating the variance and the mean of the ith AP of the jth sub-region at the time t by using a recursive Gaussian process.
7. The method of claim 1, wherein said comparing the object to be positioned received signal strength data to the region to be positioned received signal strength data set to obtain position information for the object to be positioned comprises:
Calculating the distance between the received signal strength data of the object to be positioned and the received signal strength data in the received signal strength data set of the area to be positioned;
and determining the position information of the receiving point corresponding to the received signal strength data meeting the preset distance condition as the position information of the object to be positioned.
8. the method of claim 7, wherein the distance is a Euclidean distance.
CN201811559530.4A 2018-12-19 2018-12-19 positioning method Active CN109379711B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811559530.4A CN109379711B (en) 2018-12-19 2018-12-19 positioning method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811559530.4A CN109379711B (en) 2018-12-19 2018-12-19 positioning method

Publications (2)

Publication Number Publication Date
CN109379711A CN109379711A (en) 2019-02-22
CN109379711B true CN109379711B (en) 2019-12-13

Family

ID=65370990

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811559530.4A Active CN109379711B (en) 2018-12-19 2018-12-19 positioning method

Country Status (1)

Country Link
CN (1) CN109379711B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111954158A (en) * 2020-07-01 2020-11-17 珠海高凌信息科技股份有限公司 RSS map-based joint filtering indoor single-target tracking method, device and medium
CN113344212B (en) * 2021-05-14 2024-01-19 香港中文大学（深圳） Model training method, device, computer equipment and readable storage medium
CN114430530B (en) * 2022-04-06 2022-06-21 香港中文大学（深圳） Space division method, apparatus, device, medium, and program product
CN115426712A (en) * 2022-08-25 2022-12-02 浙江工业大学 Wifi accurate robust indoor positioning method based on deep learning
CN115327478B (en) * 2022-10-10 2023-01-03 广东省电信规划设计院有限公司 Equipment positioning method and system based on wireless access point DOA estimation

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105636201A (en) * 2016-03-14 2016-06-01 中国人民解放军国防科学技术大学 Indoor positioning method based on sparse signal fingerprint database
CN108616836A (en) * 2018-04-13 2018-10-02 重庆邮电大学 A kind of WLAN positioning network-building methods based on signal statistics distribution

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
TW201042279A (en) * 2009-05-19 2010-12-01 Ralink Technology Corp Method and apparatus of using soft information for enhancing accuracy of position location estimation for a wireless communication system

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105636201A (en) * 2016-03-14 2016-06-01 中国人民解放军国防科学技术大学 Indoor positioning method based on sparse signal fingerprint database
CN108616836A (en) * 2018-04-13 2018-10-02 重庆邮电大学 A kind of WLAN positioning network-building methods based on signal statistics distribution

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
一种基于位置指纹的室内WLAN定位方法;张维，李文涛;《信息技术》;20130731;正文第0-1节 *

Also Published As

Publication number Publication date
CN109379711A (en) 2019-02-22

Similar Documents

Publication Publication Date Title
CN109379711B (en) positioning method
CN109275095B (en) Bluetooth-based indoor positioning system, positioning equipment and positioning method
CN106912105B (en) Three-dimensional positioning method based on PSO _ BP neural network
US9661473B1 (en) Methods and apparatus for determining locations of devices in confined spaces
CN112533163B (en) Indoor positioning method based on NB-IoT (NB-IoT) improved fusion ultra-wideband and Bluetooth
JP2004215258A (en) Positioning method of radio communication device using selective fusion
Aranda et al. Performance analysis of fingerprinting indoor positioning methods with BLE
CN110850363B (en) Method for carrying out dynamic filtering optimization based on real-time positioning track data
Dong et al. Dealing with insufficient location fingerprints in Wi-Fi based indoor location fingerprinting
Zhao et al. An efficient radio map updating algorithm based on K-means and Gaussian process regression
Yin et al. A novel distributed sensor fusion algorithm for RSSI-based location estimation using the unscented kalman filter
CN112526572B (en) Network switching method and positioning system for indoor and outdoor seamless navigation
CN111444467B (en) Method for local linear interpolation and prediction based on real-time positioning track data
Wu et al. Indoor positioning based on tightly coupling of PDR and one single Wi-Fi FTM AP
Munadhil et al. Distance estimation-based PSO between patient with Alzheimer’s disease and beacon node in wireless sensor networks
Elbakly et al. Cone: Zero-calibration accurate confidence estimation for indoor localization systems
İlçi et al. An investigation of different Wi-Fi signal behaviours and their effects on indoor positioning accuracy
CN114185003B (en) Indoor positioning method and system based on multi-source signal fusion
Kiring et al. Wi-Fi radio map interpolation with sparse and correlated received signal strength measurements for indoor positioning
Mo et al. A novel technique for human traffic based radio map updating in Wi-Fi indoor positioning systems
Durgaprasadarao et al. Design of deer hunting optimization algorithm for accurate 3D indoor node localization
Hosseini et al. NSGA-II based optimal Wi-Fi access point placement for indoor positioning: A BIM-based RSS prediction
Li et al. Edge computing-enabled green multisource fusion indoor positioning algorithm based on adaptive particle filter
Li et al. Dividing-and-kriging method for wireless RSS fingerprint based indoor localization
WO2022143711A1 (en) Data processing method and apparatus

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