CN113923773A - Deep learning-based indoor positioning method and device - Google Patents

Deep learning-based indoor positioning method and device Download PDF

Info

Publication number
CN113923773A
CN113923773A CN202111337466.7A CN202111337466A CN113923773A CN 113923773 A CN113923773 A CN 113923773A CN 202111337466 A CN202111337466 A CN 202111337466A CN 113923773 A CN113923773 A CN 113923773A
Authority
CN
China
Prior art keywords
information
reference point
data
characteristic value
point
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.)
Pending
Application number
CN202111337466.7A
Other languages
Chinese (zh)
Inventor
张沛昌
马琪坤
黄磊
李强
钱恭斌
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 University
Original Assignee
Shenzhen University
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 University filed Critical Shenzhen University
Priority to CN202111337466.7A priority Critical patent/CN113923773A/en
Publication of CN113923773A publication Critical patent/CN113923773A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24133—Distances to prototypes
    • G06F18/24137—Distances to cluster centroïds
    • G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G06F18/24155—Bayesian classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Biology (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Image Analysis (AREA)

Abstract

The invention provides an indoor positioning method and device based on deep learning, wherein the method comprises an off-line stage and an on-line stage; the off-line phase comprises: collecting the characteristic value of each reference point: forming a fingerprint database by using CSI information, RSS information, TDOA information, FDOA information and magnetic field information; data normalization, namely converting the data of the fingerprint database into dimensionless data and processing the dimensionless data through a multilayer CNN network to obtain CNN model parameter values of each reference point and obtain a trained CNN model; the online phase comprises: and (3) acquiring the characteristic value of each test point on line, inputting the characteristic value into the trained CNN model, calculating the probability that the target point is positioned at each reference point through a radial basis function and Bayesian classification, and then performing weighted average on the coordinates of the reference points to obtain the coordinates of the target position. The invention can improve the indoor positioning precision and the anti-interference capability.

Description

Deep learning-based indoor positioning method and device
Technical Field
The present invention relates to an indoor positioning method and device, and more particularly, to an indoor positioning method and device based on deep learning.
Background
The fifth generation communication system (5 generation, 5G) is an important information infrastructure for implementing the strategies of the network and the strong country in China, and is a high place for developing a new generation of information communication technology. The development of the internet of things is greatly promoted by the appearance of 5G, with the continuous abundance of material life of people, the demand of indoor positioning is increasing, and application scenes are also becoming more and more extensive, such as an auxiliary vehicle searching system in a parking lot, a safety tracking system of an underground tunnel, various common household robots and the like. The traditional estimation and calculation method faces the problems of more application constraint conditions and occupation of a large amount of single computing resources, deviates from the development trend of multi-scene and distributed calculation at present, and is not applicable any more. Research on a positioning algorithm which can be applied to various scenes and can be deployed to edge equipment for solution is an urgent need for further development of indoor positioning technology. In indoor positioning, conventional ranging methods such as TOA, TDOA, etc. are more or less interfered to generate noise due to a complicated indoor environment; in the position solution process, we assume that the PDF of these noises is gaussian but in practice it is far from being so. In addition, the algorithm complexity of the traditional calculation method is generally higher, so that the requirement on hardware resources is higher.
RSS is commonly used in indoor positioning systems as a readily available signal feature. However, since RSS is coarse-grained information, it is often affected by multipath effects and noise signals, and the positioning performance is not stable. In recent years, commercial WiFi devices (such as Intel 5300 wireless network cards) have begun to support CSI acquisition by the physical layer. The CSI can represent signals with finer granularity, and by analyzing the transmission conditions of different sub-channel signals respectively, the CSI can avoid the influence of multipath effect and noise as much as possible. CSI opens up new space for WiFi-based indoor positioning technology and is therefore of interest to a great deal of researchers. Most current methods based on CSI fingerprints reduce the computation resources required for positioning, resulting in low positioning accuracy.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: an indoor positioning method and device based on deep learning are provided, and the purpose is to improve indoor positioning accuracy and anti-interference capability.
In order to solve the technical problems, the invention adopts the technical scheme that: an indoor positioning method based on deep learning comprises an off-line stage and an on-line stage; the off-line phase comprises:
collecting reference point characteristic values, wherein the characteristic value of each reference point is collected: forming a fingerprint database by using CSI information, RSS information, TDOA information, FDOA information and magnetic field information;
data normalization, namely converting CSI information, RSS information, TDOA information, FDOA information and magnetic field information of each reference point in the fingerprint database into dimensionless data;
model training, namely processing the dimensionless data of each reference point through a multilayer CNN network to obtain CNN model parameter values of each reference point and obtain a trained CNN model;
the online phase comprises:
collecting characteristic values of the test points, and collecting the characteristic value of each test point on line: CSI information, RSS information, TDOA information, FDOA information, and magnetic field information;
fingerprint matching, and characteristic values of online collected test points: the CSI information, the RSS information, the TDOA information, the FDOA information and the magnetic field information are input into the trained CNN model, the probability that a target point is located at each reference point is calculated through radial basis functions and Bayesian classification, and then the coordinates of the reference points are weighted and averaged to obtain the coordinates of the target position.
Further, the formula for converting the CSI information, RSS information, TDOA information, FDOA information, and magnetic field information of each reference point in the fingerprint database into dimensionless data is as follows:
Figure BDA0003348999120000021
wherein the data represents each data in the fingerprint database, and the data' is normalized sampling data.
Further, the fingerprint matching adopts a probability algorithm to match the fingerprint database, and the probability algorithm is as follows:
Figure BDA0003348999120000022
Pris the probability of a test point being at i of a sampling point, NcRepresenting the number of reference points, ciIs the position coordinate of the reference point i in the fingerprint library, Pr(ci) Is that the target is located at the reference point ciT is the input characteristic value data.
Further, the radial basis function is:
Figure BDA0003348999120000031
where T is the input characteristic value data, Pr(T|ci) To be at a reference point ciOutput Y oftSimilarity to itself, λTσTRespectively, the variance and variance parameters of the input characteristic value data T.
Further, the formula for performing weighted average on the coordinates of the reference point is as follows:
Figure BDA0003348999120000032
wherein the content of the first and second substances,
Figure BDA0003348999120000033
coordinates representing the target position, NcRepresenting the number of reference points, ciIs the location coordinate of the reference point i in the fingerprint library.
The other technical scheme of the invention is as follows: an indoor positioning device based on deep learning comprises an offline module and an online module; the offline module includes:
a reference point characteristic value acquisition unit, configured to acquire a characteristic value of each reference point: forming a fingerprint database by using CSI information, RSS information, TDOA information, FDOA information and magnetic field information;
a data normalization unit for converting the CSI information, RSS information, TDOA information, FDOA information, and magnetic field information of each reference point in the fingerprint database into dimensionless data;
the model training unit is used for processing the dimensionless data of each reference point through the multilayer CNN network to obtain the CNN model parameter value of each reference point and obtain a trained CNN model;
the online module includes:
the test point characteristic value acquisition unit is used for acquiring the characteristic value of each test point on line: CSI information, RSS information, TDOA information, FDOA information, and magnetic field information;
the fingerprint matching unit is used for acquiring the characteristic values of the test points on line: the CSI information, the RSS information, the TDOA information, the FDOA information and the magnetic field information are input into the trained CNN model, the probability that a target point is located at each reference point is calculated through radial basis functions and Bayesian classification, and then the coordinates of the reference points are weighted and averaged to obtain the coordinates of the target position.
Further, in the data normalization unit, the formula for converting the CSI information, RSS information, TDOA information, FDOA information, and magnetic field information of each reference point in the fingerprint database into dimensionless data is as follows:
Figure BDA0003348999120000034
wherein the data represents each data in the fingerprint database, and the data' is normalized sampling data.
Further, in the fingerprint matching unit, the fingerprint matching adopts a probability algorithm to match the fingerprint database, and the probability algorithm is as follows:
Figure BDA0003348999120000041
Pris the probability of a test point being at i of a sampling point, NcRepresenting the number of reference points, ciIs the position coordinate of the reference point i in the fingerprint library, Pr(ci) Is that the target is located at the reference point ciT is the input characteristic value data.
Further, in the fingerprint matching unit, the radial basis functions used are:
Figure BDA0003348999120000042
where T is the input characteristic value data, Pr(T|ci) To be at a reference point ciOutput Y oftSimilarity to itself, λTσTRespectively, the variance and variance parameters of the input characteristic value data T.
Further, in the fingerprint matching unit, a formula for performing weighted average on the coordinates of the reference point is as follows:
Figure BDA0003348999120000043
wherein the content of the first and second substances,
Figure BDA0003348999120000044
coordinates representing the target position, NcRepresenting the number of reference points, ciIs the location coordinate of the reference point i in the fingerprint library.
The invention has the beneficial effects that: firstly, after collecting CSI information, RSS information, TDOA information, FDOA information and magnetic field information, processing the data by using a CNN network; in the processing process, original measurement data are processed through a multi-layer CNN network, CNN model parameter values at each reference point are used as fingerprints, the characteristics of the CNN model are more representative, and measurement errors and interference which may exist when the original data are directly used for the fingerprints are also avoided; then in the fingerprint matching stage, the probability that the target is located at each reference point is calculated through a radial basis function and Bayesian law, and then the positioning is realized through the weighted average of the coordinates of the reference points. The CSI information, the RSS information, the TDOA information, the FDOA information and the magnetic field information are collected to be used as characteristic values of positioning, and the accuracy and the anti-interference capability of indoor positioning can be improved.
Drawings
The following detailed description of the invention refers to the accompanying drawings.
Fig. 1 is a flowchart of an indoor positioning method based on deep learning according to an embodiment of the present invention;
FIG. 2 is a block diagram of an indoor positioning apparatus based on deep learning according to an embodiment of the present invention;
FIG. 3 is a schematic block diagram of a computer device in accordance with a specific embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should be further understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.
As shown in fig. 1, the first embodiment of the present invention is: an indoor positioning method based on deep learning comprises an off-line stage and an on-line stage; the off-line phase comprises:
s11, collecting characteristic values of reference points, wherein the characteristic value of each reference point is collected: forming a fingerprint database by using CSI information, RSS information, TDOA information, FDOA information and magnetic field information;
s12, data normalization, namely converting CSI information, RSS information, TDOA information, FDOA information and magnetic field information of each reference point in the fingerprint database into dimensionless data;
s13, model training, namely processing the dimensionless data of each reference point through a multilayer CNN network to obtain CNN model parameter values of each reference point and obtain a trained CNN model;
the online phase comprises:
s21, collecting characteristic values of the test points, and collecting the characteristic value of each test point on line: CSI information, RSS information, TDOA information, FDOA information, and magnetic field information;
s22, fingerprint matching, and feature values of online collected test points: the CSI information, the RSS information, the TDOA information, the FDOA information and the magnetic field information are input into the trained CNN model, the probability that a target point is located at each reference point is calculated through radial basis functions and Bayesian classification, and then the coordinates of the reference points are weighted and averaged to obtain the coordinates of the target position.
In the embodiment, Channel State Information (CSI), Magnetic Field Strength (MFS), Received Signal Strength (RSS) of WiFi, Time Difference of Arrival (TDOA) measurement value after clock synchronization, and (Frequency Difference of Arrival, FDOA) are fused by deep learning to realize high-precision indoor positioning, and the above information alone can be used as a position feature, but independence cannot be guaranteed in a large-area indoor environment.
In step S12, the formula for converting the CSI information, RSS information, TDOA information, FDOA information, and magnetic field information of each reference point in the fingerprint database into dimensionless data is as follows:
Figure BDA0003348999120000061
wherein the data represents each data in the fingerprint database, and the data' is normalized sampling data.
In step S22, the fingerprint matching uses a probability algorithm to match the fingerprint database, where the probability algorithm is:
Figure BDA0003348999120000062
Pris the probability of a test point being at i of a sampling point, NcRepresenting the number of reference points, ciIs the position coordinate of the reference point i in the fingerprint library, Pr(ci) Is that the target is located at the reference point ciT is the input characteristic value data. It can be assumed that Pr(ci) Subject to uniform distribution, then
Figure BDA0003348999120000063
Wherein the radial basis function is:
Figure BDA0003348999120000071
where T is the input characteristic value data, Pr(T|ci) To be at a reference point ciOutput Y oftSimilarity to itself，λTσTRespectively, the variance and variance parameters of the input characteristic value data T.
The formula for performing weighted average on the reference point coordinates is as follows:
Figure BDA0003348999120000072
wherein the content of the first and second substances,
Figure BDA0003348999120000073
coordinates representing the target position, NcRepresenting the number of reference points, ciIs the location coordinate of the reference point i in the fingerprint library.
By the method, the traditional iterative process is converted into a primary matching problem, algorithm complexity is greatly reduced, the trained network model can be deployed to edge equipment, the CNN network can be vertically split along with continuous development of subsequent technologies, each layer of the network is deployed to a server, generated weight values can be interacted in a JSON format in a short connection mode, and caching can be directly carried out by using a NoSQL technology.
As shown in fig. 2, another embodiment of the present invention is: an indoor positioning device based on deep learning comprises an offline module and an online module; the offline module includes:
a reference point characteristic value acquisition unit 11, configured to acquire a characteristic value of each reference point: forming a fingerprint database by using CSI information, RSS information, TDOA information, FDOA information and magnetic field information;
a data normalization unit 12 for converting CSI information, RSS information, TDOA information, FDOA information, and magnetic field information of each reference point in the fingerprint database into dimensionless data;
the model training unit 13 is configured to process the dimensionless data of each reference point through the multilayer CNN network to obtain a CNN model parameter value of each reference point, so as to obtain a trained CNN model;
the online module includes:
a test point characteristic value acquisition unit 21, configured to acquire a characteristic value of each test point on line: CSI information, RSS information, TDOA information, FDOA information, and magnetic field information;
a fingerprint matching unit 22, configured to: the CSI information, the RSS information, the TDOA information, the FDOA information and the magnetic field information are input into the trained CNN model, the probability that a target point is located at each reference point is calculated through radial basis functions and Bayesian classification, and then the coordinates of the reference points are weighted and averaged to obtain the coordinates of the target position.
In the data normalization unit 12, the formula for converting the CSI information, RSS information, TDOA information, FDOA information, and magnetic field information of each reference point in the fingerprint database into dimensionless data is as follows:
Figure BDA0003348999120000081
wherein the data represents each data in the fingerprint database, and the data' is normalized sampling data.
In the fingerprint matching unit 22, the fingerprint matching adopts a probability algorithm to match the fingerprint database, where the probability algorithm is:
Figure BDA0003348999120000082
Pris the probability of a test point being at i of a sampling point, NCRepresenting the number of reference points, ciIs the position coordinate of the reference point i in the fingerprint library, Pr(ci) Is that the target is located at the reference point ciT is the input characteristic value data.
Among these, in the fingerprint matching unit 22, the radial basis functions used are:
Figure BDA0003348999120000083
wherein T is inputCharacteristic value data, Pr(T|ci) To be at a reference point ciOutput Y oftSimilarity to itself, λTσTRespectively, the variance and variance parameters of the input characteristic value data T.
In the fingerprint matching unit 22, the formula for performing weighted average on the coordinates of the reference point is as follows:
Figure BDA0003348999120000084
wherein the content of the first and second substances,
Figure BDA0003348999120000085
coordinates representing the target position, NCRepresenting the number of reference points, ciIs the location coordinate of the reference point i in the fingerprint library.
It should be noted that, as can be clearly understood by those skilled in the art, the detailed implementation process of the indoor positioning device and each unit based on deep learning may refer to the corresponding description in the foregoing method embodiment, and for convenience and brevity of description, no further description is provided herein.
The above-described deep learning based indoor positioning apparatus may be implemented in the form of a computer program that can be run on a computer device as shown in fig. 3.
Referring to fig. 3, fig. 3 is a schematic block diagram of a computer device according to an embodiment of the present application. The computer device 500 may be a terminal or a server, where the terminal may be an electronic device with a communication function, such as a smart phone, a tablet computer, a notebook computer, a desktop computer, a personal digital assistant, and a wearable device. The server may be an independent server or a server cluster composed of a plurality of servers.
Referring to fig. 3, the computer device 500 includes a processor 502, memory, and a network interface 505 connected by a system bus 501, where the memory may include a non-volatile storage medium 503 and an internal memory 504.
The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 comprises program instructions that, when executed, cause the processor 502 to perform a deep learning based indoor positioning method.
The processor 502 is used to provide computing and control capabilities to support the operation of the overall computer device 500.
The internal memory 504 provides an environment for running the computer program 5032 in the non-volatile storage medium 503, and when the computer program 5032 is executed by the processor 502, the processor 502 can be enabled to execute a deep learning based indoor positioning method.
The network interface 505 is used for network communication with other devices. Those skilled in the art will appreciate that the configuration shown in fig. 3 is a block diagram of only a portion of the configuration associated with the present application and does not constitute a limitation of the computer device 500 to which the present application may be applied, and that a particular computer device 500 may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
Wherein the processor 502 is configured to run a computer program 5032 stored in the memory to implement the deep learning based indoor positioning method as above.
It should be understood that in the embodiment of the present Application, the Processor 502 may be a Central Processing Unit (CPU), and the Processor 502 may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, and the like. Wherein a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
It will be understood by those skilled in the art that all or part of the flow of the method implementing the above embodiments may be implemented by a computer program instructing associated hardware. The computer program includes program instructions, and the computer program may be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the flow steps of the embodiments of the method described above.
Accordingly, the present invention also provides a storage medium. The storage medium may be a computer-readable storage medium. The storage medium stores a computer program, wherein the computer program comprises program instructions. The program instructions, when executed by the processor, cause the processor to perform the deep learning based indoor positioning method as described above.
The storage medium may be a usb disk, a removable hard disk, a Read-Only Memory (ROM), a magnetic disk, or an optical disk, which can store various computer readable storage media.
Those of ordinary skill in the art will appreciate that the elements and algorithm steps of the examples described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and that the components and steps of the examples have been described in a functional general in the foregoing description for the purpose of illustrating clearly the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative. For example, the division of each unit is only one logic function division, and there may be another division manner in actual implementation. For example, various elements or components may be combined or may be integrated into another system, or some features may be omitted, or not implemented.
The steps in the method of the embodiment of the invention can be sequentially adjusted, combined and deleted according to actual needs. The units in the device of the embodiment of the invention can be merged, divided and deleted according to actual needs. In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a storage medium. Based on such understanding, the technical solution of the present invention essentially or partially contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a terminal, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims (10)

1. An indoor positioning method based on deep learning is characterized in that: comprises an off-line stage and an on-line stage; the off-line phase comprises:
collecting reference point characteristic values, wherein the characteristic value of each reference point is collected: forming a fingerprint database by using CSI information, RSS information, TD0A information, FDOA information and magnetic field information;
data normalization, namely converting CSI information, RSS information, TDOA information, FDOA information and magnetic field information of each reference point in the fingerprint database into dimensionless data;
model training, namely processing the dimensionless data of each reference point through a multilayer CNN network to obtain CNN model parameter values of each reference point and obtain a trained CNN model;
the online phase comprises:
collecting characteristic values of the test points, and collecting the characteristic value of each test point on line: CSI information, RSS information, TDOA information, FDOA information, and magnetic field information;
fingerprint matching, and characteristic values of online collected test points: the CSI information, the RSS information, the TDOA information, the FDOA information and the magnetic field information are input into the trained CNN model, the probability that a target point is located at each reference point is calculated through radial basis functions and Bayesian classification, and then the coordinates of the reference points are weighted and averaged to obtain the coordinates of the target position.
2. The deep learning-based indoor positioning method of claim 1, wherein: the formula for converting the CSI information, RSS information, TDOA information, FDOA information, and magnetic field information of each reference point in the fingerprint database into dimensionless data is as follows:
Figure FDA0003348999110000011
wherein the data represents each data in the fingerprint database, and the data' is normalized sampling data.
3. The deep learning-based indoor positioning method of claim 1, wherein: the fingerprint matching adopts a probability algorithm to match the fingerprint database, and the probability algorithm is as follows:
Figure FDA0003348999110000012
Pris the probability of a test point being at i of a sampling point, NCRepresenting the number of reference points, ciIs the position coordinate of the reference point i in the fingerprint library, Pr(ci) Is that the target is located at the reference point ciA priori probability of (A), T being the characteristic value data of the input。
4. The deep learning-based indoor positioning method of claim 3, wherein: the radial basis function is:
Figure FDA0003348999110000021
where T is the input characteristic value data, Pr(T|ci) To be at a reference point ciOutput Y oftSimilarity to itself, λTσTRespectively, the variance and variance parameters of the input characteristic value data T.
5. The deep learning-based indoor positioning method of claim 4, wherein: the formula for weighted averaging of the reference point coordinates is:
Figure FDA0003348999110000022
wherein the content of the first and second substances,
Figure FDA0003348999110000023
coordinates representing the target position, NCRepresenting the number of reference points, ciIs the location coordinate of the reference point i in the fingerprint library.
6. The utility model provides an indoor positioner based on degree of depth study which characterized in that: the system comprises an offline module and an online module; the offline module includes:
a reference point characteristic value acquisition unit, configured to acquire a characteristic value of each reference point: forming a fingerprint database by using CSI information, RSS information, TDOA information, FDOA information and magnetic field information;
a data normalization unit for converting the CSI information, RSS information, TDOA information, FDOA information, and magnetic field information of each reference point in the fingerprint database into dimensionless data;
the model training unit is used for processing the dimensionless data of each reference point through the multilayer CNN network to obtain the CNN model parameter value of each reference point and obtain a trained CNN model;
the online module includes:
the test point characteristic value acquisition unit is used for acquiring the characteristic value of each test point on line: CSI information, RSS information, TDOA information, FDOA information, and magnetic field information;
the fingerprint matching unit is used for acquiring the characteristic values of the test points on line: the CSI information, the RSS information, the TDOA information, the FDOA information and the magnetic field information are input into the trained CNN model, the probability that a target point is located at each reference point is calculated through radial basis functions and Bayesian classification, and then the coordinates of the reference points are weighted and averaged to obtain the coordinates of the target position.
7. The deep learning based indoor positioning apparatus of claim 6, wherein: in the data normalization unit, the formula for converting the CSI information, RSS information, TDOA information, FDOA information, and magnetic field information of each reference point in the fingerprint database into dimensionless data is as follows:
Figure FDA0003348999110000031
wherein the data represents each data in the fingerprint database, and the data' is normalized sampling data.
8. The deep learning based indoor positioning apparatus of claim 6, wherein: in the fingerprint matching unit, fingerprint matching adopts a probability algorithm to match a fingerprint database, and the probability algorithm is as follows:
Figure FDA0003348999110000032
Prfor test points located at the samplesProbability of i of a point, NCRepresenting the number of reference points, ciIs the position coordinate of the reference point i in the fingerprint library, Pr(ci) Is that the target is located at the reference point ciT is the input characteristic value data.
9. The deep learning based indoor positioning apparatus of claim 8, wherein: in the fingerprint matching unit, the radial basis functions used are:
Figure FDA0003348999110000033
where T is the input characteristic value data, Pr(T|ci) To be at a reference point ciOutput Y oftSimilarity to itself, λTσTRespectively, the variance and variance parameters of the input characteristic value data T.
10. The deep learning based indoor positioning apparatus of claim 9, wherein: in the fingerprint matching unit, the formula for performing weighted average on the coordinates of the reference points is as follows:
Figure FDA0003348999110000034
wherein the content of the first and second substances,
Figure FDA0003348999110000035
coordinates representing the target position, NCRepresenting the number of reference points, ciIs the location coordinate of the reference point i in the fingerprint library.
CN202111337466.7A 2021-11-11 2021-11-11 Deep learning-based indoor positioning method and device Pending CN113923773A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111337466.7A CN113923773A (en) 2021-11-11 2021-11-11 Deep learning-based indoor positioning method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111337466.7A CN113923773A (en) 2021-11-11 2021-11-11 Deep learning-based indoor positioning method and device

Publications (1)

Publication Number Publication Date
CN113923773A true CN113923773A (en) 2022-01-11

Family

ID=79246065

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111337466.7A Pending CN113923773A (en) 2021-11-11 2021-11-11 Deep learning-based indoor positioning method and device

Country Status (1)

Country Link
CN (1) CN113923773A (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108696932A (en) * 2018-04-09 2018-10-23 西安交通大学 It is a kind of using CSI multipaths and the outdoor fingerprint positioning method of machine learning
CN110381440A (en) * 2019-06-16 2019-10-25 西安电子科技大学 The fingerprint indoor orientation method of joint RSS and CSI based on deep learning
US20210105578A1 (en) * 2019-10-04 2021-04-08 Samsung Electronics Co., Ltd. System and method for wifi-based indoor localization via unsupervised domain adaptation

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108696932A (en) * 2018-04-09 2018-10-23 西安交通大学 It is a kind of using CSI multipaths and the outdoor fingerprint positioning method of machine learning
CN110381440A (en) * 2019-06-16 2019-10-25 西安电子科技大学 The fingerprint indoor orientation method of joint RSS and CSI based on deep learning
US20210105578A1 (en) * 2019-10-04 2021-04-08 Samsung Electronics Co., Ltd. System and method for wifi-based indoor localization via unsupervised domain adaptation

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
王英等: "基于卷积神经网络的室内指纹定位算法", 《软件学报》, pages 63 - 72 *

Similar Documents

Publication Publication Date Title
US9268399B2 (en) Adaptive sensor sampling for power efficient context aware inferences
CN107703480A (en) Mixed kernel function indoor orientation method based on machine learning
US20120054261A1 (en) Dual modeling environment
CN111191555B (en) Target tracking method, medium and system combining high-low spatial frequency characteristics
CN111738319B (en) Clustering result evaluation method and device based on large-scale samples
CN110057355B (en) Indoor positioning method, device and system and computing equipment
Chen et al. A wifi indoor localization method based on dilated cnn and support vector regression
CN113657249B (en) Training method, prediction method, device, electronic equipment and storage medium
CN114328785A (en) Method and device for extracting road information
CN113923773A (en) Deep learning-based indoor positioning method and device
CN111241106A (en) Approximate data processing method, device, medium and electronic equipment
CN115049590B (en) Image processing method and device, electronic equipment and storage medium
WO2023193395A1 (en) Mixed-frequency data imputation model creation method based on functional data analysis
CN113065585B (en) Training method and device of image synthesis model and electronic equipment
CN113189541B (en) Positioning method, device and equipment
CN112580701B (en) Mean value estimation method and device based on classification transformation disturbance mechanism
Hai-Jun et al. Fuzzy entropy clustering using possibilistic approach
WO2021120489A1 (en) Adaptive fault tolerance processing method and system for spatial information data
Zhang et al. Hierarchical Image Retrieval Method Based on Bag-of-Visual-Word and Eight-point Algorithm with Feature Clouds for Visual Indoor Positioning
CN113011597A (en) Deep learning method and device for regression task
CN107567093B (en) Fingerprint positioning method based on robust sparse representation clustering
CN113095353B (en) Underdetermined blind source separation method based on AP clustering
CN111160969A (en) Power price prediction method and device
CN117493775B (en) Relative navigation method and device of data chain, electronic equipment and storage medium
CN116402842B (en) Edge defect detection method, device, computer equipment and storage medium

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20220111