CN109444813B - RFID indoor positioning method based on BP and DNN double neural networks - Google Patents

RFID indoor positioning method based on BP and DNN double neural networks Download PDF

Info

Publication number
CN109444813B
CN109444813B CN201811255127.2A CN201811255127A CN109444813B CN 109444813 B CN109444813 B CN 109444813B CN 201811255127 A CN201811255127 A CN 201811255127A CN 109444813 B CN109444813 B CN 109444813B
Authority
CN
China
Prior art keywords
rssi
path loss
dnn
loss coefficient
neural network
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
CN201811255127.2A
Other languages
Chinese (zh)
Other versions
CN109444813A (en
Inventor
叶宁
陈龙鹏
徐康
王娟
黄海平
程晶晶
林巧民
王汝传
凌鑫元
贾成栋
马铭辰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University of Posts and Telecommunications
Original Assignee
Nanjing University of Posts and Telecommunications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University of Posts and Telecommunications filed Critical Nanjing University of Posts and Telecommunications
Priority to CN201811255127.2A priority Critical patent/CN109444813B/en
Publication of CN109444813A publication Critical patent/CN109444813A/en
Application granted granted Critical
Publication of CN109444813B publication Critical patent/CN109444813B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
    • G01S5/02—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves
    • G01S5/0278—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using radio waves involving statistical or probabilistic considerations
    • 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
    • G06N3/084—Backpropagation, e.g. using gradient descent

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Probability & Statistics with Applications (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Position Fixing By Use Of Radio Waves (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

An RFID indoor positioning method based on BP and DNN double neural networks overcomes the defect that the path loss coefficient n is set to be constant in the same environment in the traditional RSSI-based indoor positioning technology, establishes a conversion model of signal intensity and the path loss coefficient by combining the neural network technology, accurately predicts the path loss coefficient n at different positions, reduces the error caused by the fixation of the path loss coefficient n in the traditional RSSI-based positioning method, and improves the positioning precision of a system; by combining the BP network and the deep neural network, the path loss coefficients n of different positions output by the BP network and the received signal strength of the tag to be tested are used as the input of the DNN, and the corresponding path loss coefficients n can be output according to different environments, so that the coordinates of the tag to be tested can be predicted more accurately, and the robustness of the system is improved; the coordinates of the label to be detected are output by combining with the deep neural network, the label to be detected can be positioned in real time, and the defect that the traditional positioning method is poor in real-time performance is overcome.

Description

RFID indoor positioning method based on BP and DNN double neural networks
Technical Field
The invention belongs to the technical field of wireless communication positioning, and particularly provides a BP and DNN dual neural network-based RFID indoor positioning method which is used in an indoor environment and improves positioning accuracy by combining a neural network technology.
Background
The research at home and abroad combines the characteristics of indoor channel environment and the existing mature wireless technology to provide a plurality of solutions. According to different sensing environment parameters and data acquisition modes, indoor positioning can be divided into seven categories, namely indoor positioning based on a WiFi technology, a Bluetooth technology, a ZigBee technology, an ultra-wideband technology, an ultrasonic technology, an infrared technology and an RFID technology. Radio Frequency Identification (RFID) is a technology that automatically identifies a target signal object and acquires related information using a Radio Frequency signal. Due to the fact that the RFID technology has non-contact and non-line-of-sight functions, manual intervention is not needed in identification, a plurality of labels can be identified at the same time, and meanwhile the RFID technology has the advantages of being fast and convenient to operate, high in positioning accuracy and the like, and becomes an optimal indoor positioning technology.
There are two kinds of indoor positioning methods based on RFID, which are a ranging method and a non-ranging method. The location method based on the ranging mainly comprises four methods of time of arrival (TOA), time difference of arrival (TDOA), angle of arrival (AOA) and Received Signal Strength (RSSI), and the location method based on the non-ranging comprises scene analysis location and proximity location. The positioning method based on the RSSI is simple in distance measurement and is the most commonly used indoor positioning method, but the positioning method has the problems of low positioning accuracy, large environmental influence and the like.
Disclosure of Invention
The invention provides an RFID positioning method combined with a neural network technology, which solves the problem that the complicated indoor environment influences positioning in the existing indoor positioning technology based on RFID, improves the positioning precision and overcomes the problem that the traditional positioning algorithm has poor real-time performance.
An RFID indoor positioning method based on BP and DNN dual neural networks is characterized by comprising the following steps: comprises the following steps:
the method comprises the following steps: arranging readers and RFID reference tags prepared in advance according to the actual situation of an indoor positioning area according to a certain rule, measuring the distance from the reference tags to the readers, recording the signal strength value RSSI of each reader for receiving the RFID reference tags and the coordinates of the positions of the tags, and obtaining an original training data set;
step two: carrying out denoising pretreatment on the acquired signal strength value RSSI; in order to reduce interference and eliminate errors caused by small probability events, RSSI values in a large probability interval are selected as effective data through a Gaussian model, and then the arithmetic mean value of the RSSI values is obtained as the output of filtering;
step three: substituting the measured distance from the reference label to the reader-writer and the RSSI value received by the reader-writer into a distance measurement formula to calculate a path loss coefficient n;
step four: concentrating the filtered RSSI and the path loss coefficient n together to construct a new training set of a BP neural network model, concentrating the RSSI, the path loss coefficient n and the coordinates of a label to be detected together to construct a new training set of a DNN model;
step five: training a BP neural network; the BP network has a three-layer structure, the filtered RSSI value is used as the input of the BP network, and the path loss coefficient n i As output, training a BP network model;
step six: training a deep neural network; the filtered RSSI value and the environmental loss coefficient n output by the BP network i The actual coordinates of the reference labels are used as output, and a deep neural network model is trained;
step seven: filtering the RSSI received by the reader, and inputting the filtered RSSI into a BP network model to predict a path loss coefficient n i And inputting the measured data and the RSSI into a DNN model, and predicting the coordinates of the to-be-measured label in real time on line.
Further, in the second step, the specific denoising processing step is as follows:
first, the RSSI values obey (0, σ) 2 ) The probability density function of the gaussian distribution of (1) is shown as:
Figure BDA0001842542550000031
wherein the content of the first and second substances,
Figure BDA0001842542550000032
Figure BDA0001842542550000033
then interval (μ - σ < RSSI k < μ + σ) is as follows:
P(μ-σ＜RSSI k ＜μ+σ)＝F(μ+σ)-F(μ-σ)＝φ(1)-φ(-1)＝0.6826
the arithmetic mean of these RSSI values is calculated:
Figure BDA0001842542550000034
wherein, N is the number of signal strength values within the maximum probability interval in k consecutive measurements.
Further, in the third step, the ranging formula is as follows:
RSSI(d)＝RSSI(d 0 )-10nlgd
further, in the fifth step, because it is very difficult to acquire a large amount of data in an actual situation, a k-fold cross validation method is adopted for training, the acquired data set D is divided into k mutually exclusive subsets with similar sizes, one subset is reserved as a test set, the other k-1 subsets are used as training sets, cross validation is repeated k times, and an average value of k test results is returned.
Compared with the prior art, the method has the following beneficial effects:
(1) The method provided by the invention overcomes the defect that the path loss coefficient n is set as a constant in the same environment in the traditional RSSI-based indoor positioning technology, establishes a conversion model of the signal intensity and the path loss coefficient by combining a neural network technology, accurately predicts the path loss coefficient n at different positions, reduces the error generated by the fixation of the path loss coefficient n in the traditional RSSI-based positioning method, and improves the positioning precision of the system.
(2) The method provided by the invention combines the BP network and the deep neural network, takes the path loss coefficients n of different positions output by the BP network and the received signal strength of the tag to be tested as the input of the DNN, and can output the corresponding path loss coefficients n according to different environments, thereby more accurately predicting the coordinates of the tag to be tested and improving the robustness of the system.
(3) The method provided by the invention can be used for outputting the coordinates of the label to be detected by combining with the deep neural network, can be used for positioning the label to be detected in real time, and overcomes the defect of poor real-time performance of the traditional positioning method.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention.
Fig. 2 is a top view of a room of an indoor positioning system based on the method of the present invention.
Fig. 3 is a schematic structural diagram of a BP neural network.
Fig. 4 is a schematic diagram of the structure of a deep neural network.
Detailed Description
The technical scheme of the invention is further explained in detail by combining the drawings in the specification.
The invention provides an RFID (radio frequency identification) indoor positioning method based on BP (Back propagation) and DNN (deep neural network), which is characterized in that the method combines the traditional RSSI (received signal strength indicator) -based indoor positioning method with a neural network, and fits the functional relation between the RSSI and a path loss coefficient n as well as the coordinates of a label to be detected by utilizing the characteristic that the neural network can fit any continuous functional relation. The method mainly comprises the following steps: arranging a reader and a tag in a room in an off-line stage, performing data acquisition and data denoising pretreatment to obtain a training data set, and constructing and training a BP network model and a deep neural network model; after the model is trained, the RSSI and the n input DNN model of the signal strength value received by the reader are collected on line, and finally the label coordinate is obtained.
The embodiment of the present invention will be described by taking an indoor environment shown in fig. 2 as an example.
The method comprises the following steps: firstly, environment arrangement is carried out, and data are collected. Book (notebook)The positioning system comprises a reader-writer, an RFID label and a computer terminal. Four readers are respectively arranged at four corners in a room, and reference tags are prepared to be arranged at 50 reference points uniformly arranged in the room in advance. The four readers continuously sample each reference label for 30 times respectively to obtain the signal strength value RSSI of the ith label k i，j Is denoted as R i Wherein i =1,2, 3. # 50, j =1,2,3,4, k =1,2, 3. # 30, and records the coordinates P of the corresponding i-th tag i (x i ，y i ). (the signal intensity value of each label is combined with the coordinates to obtain a raw training data set D containing noise = { (R) 1, ，P 1 ),(R 2 ，P 2 )，...(R i ，P i )}。)
Step two: and denoising and preprocessing the acquired signal strength value RSSI. In order to reduce interference and eliminate errors caused by small probability events, RSSI values in a large probability interval are selected as effective data through a Gaussian model, and the relationship between signal strength values and a Gaussian function is as follows:
Figure BDA0001842542550000061
wherein the content of the first and second substances,
Figure BDA0001842542550000062
Figure BDA0001842542550000063
RSSI k i,j for the kth signal intensity value of the ith tag read by the jth reader, i =1,2, 3., 50, j =1,2,3,4, k =1,2, 3., 30.
And (3) taking the received signal strength value in the large probability interval as effective data to be reserved, and then averaging all the effective data in k times of measurement, wherein the process is as follows:
Figure BDA0001842542550000064
wherein m is the number D 'of signal intensity values in a coincidence probability interval in k times of continuous measurement' i,j R 'represents the average value of the RSSI of the filtered signal strength values read by the jth reader-writer of the ith tag' i ＝{D’ i,1 ，D‘ i,2 ，D’ i,3 ，D‘ i，4 }。
Step three: ten distance read-write devices d after filtering 0 Signal Strength value RSSI, i.e. D ', of reference tag at =1 m' d D =1, 2.., 10, the average of which is calculated
Figure BDA0001842542550000065
The path loss coefficient n can be obtained by formula i As follows:
Figure BDA0001842542550000066
step four: an offline training data set is established. Obtaining a high-quality data set according to the second step, and putting the signal strength value RSSI and the path loss coefficient n which are obtained after filtering together to obtain a training data set D 'of the BP neural network' 1 ＝{(R’ 1 ，n 1 ),(R’ 2 ，n 2 )，...，(R’ i ，n i ) Fifthly, the filtered signal intensity value R is used i And a path loss coefficient n i And label coordinates P i Are combined to obtain a new training data set D 'of the deep neural network' 2 ＝{(R’ 1 ，n 1 ，P 1 )，(R’ 2 ，n 2, ，P 2 )，...(R’ i ，n i ，P i )}，i＝1,2,3，...，50。
Step five: and training the BP network model. After a new training data set is obtained, a k-fold cross-validation method is adopted to performAnd (5) training. Data set D' 1 And dividing the model into 10 parts, wherein one part is used as a test set, and the other 9 parts are used as training sets, and repeatedly dividing and training the model for 10 times to obtain the model with the minimum generalization error, so that different path loss coefficients n under different environments are obtained, and the influence of the environment on the positioning accuracy is avoided.
Step six: the DNN network model is trained. Similar to the training BP neural network model, a data set D 'of DNN is obtained' 2 And then, training by adopting a k-fold cross validation method, optimizing model parameters to obtain an optimal DNN model, ensuring the real-time performance of positioning by the model, and effectively avoiding positioning errors caused by complicated and variable indoor environments.
Step seven: and predicting the coordinates of the tag to be detected in real time on line. And (3) after an RFID label enters a receiving area of the reader, preprocessing data in the third step to obtain a filtered signal strength value RSSI, inputting the filtered signal strength value RSSI into a trained BP neural network model, outputting a path loss coefficient n, and inputting the filtered signal strength value RSSI and the filtered signal strength value RSSI into a DNN, so that the accurate coordinate of the label to be detected can be output.
The above description is only a preferred embodiment of the present invention, and the scope of the present invention is not limited to the above embodiment, but equivalent modifications or changes made by those skilled in the art according to the present disclosure should be included in the scope of the present invention as set forth in the appended claims.

Claims (4)

1. An RFID indoor positioning method based on BP and DNN dual neural networks is characterized in that: comprises the following steps:
the method comprises the following steps: arranging readers and RFID reference tags prepared in advance according to the actual situation of an indoor positioning area according to a certain rule, measuring the distance from the reference tags to the readers, recording the signal strength value RSSI of each reader for receiving the RFID reference tag and the coordinates of the position of the tag, and obtaining an original training data set;
step two: carrying out denoising pretreatment on the acquired signal strength value RSSI; in order to reduce interference and eliminate errors caused by small probability events, RSSI values in a large probability interval are selected as effective data through a Gaussian model, and then the arithmetic mean value of the RSSI values is obtained as the output of filtering;
step three: substituting the measured distance between the reference label and the reader-writer and the RSSI value received by the reader-writer into a distance measurement formula to calculate a path loss coefficient n;
step four: concentrating the filtered RSSI and the path loss coefficient n together to construct a new training set of a BP neural network model, concentrating the RSSI, the path loss coefficient n and the coordinates of a label to be detected together to construct a new training set of a DNN model;
step five: training a BP neural network; the BP neural network is of a three-layer structure, the RSSI value after filtering is used as the input of the BP network, the path loss coefficient n is used as the output, and a BP network model is trained;
step six: training a deep neural network; the filtered RSSI value and the environmental loss coefficient n output by the BP network i The actual coordinates of the reference labels are used as output, and a deep neural network model is trained;
step seven: and filtering the RSSI received by the reader, inputting the filtered RSSI into a BP network model to predict a path loss coefficient n, inputting the path loss coefficient n and the RSSI into a DNN model, and predicting the coordinates of the tag to be measured in real time on line.
2. The RFID indoor positioning method based on BP and DNN dual neural network as claimed in claim 1, characterized in that: in the second step, the specific denoising processing steps are as follows:
first, the RSSI values obey (0, σ) 2 ) The probability density function of the gaussian distribution of (1) is shown as:
Figure FDA0003809056490000021
wherein the content of the first and second substances,
Figure FDA0003809056490000022
Figure FDA0003809056490000023
then interval (μ - σ < RSSI) k < μ + σ) is as follows:
P(μ-σ＜RSSI k ＜μ+σ)＝F(μ+σ)-F(μ-σ)＝φ(1)-φ(-1)＝0.6826
the arithmetic mean of these RSSI values is calculated:
Figure FDA0003809056490000024
and N is the number of signal strength values in the maximum probability interval in k times of continuous measurement.
3. The RFID indoor positioning method based on BP and DNN dual neural networks as claimed in claim 1, characterized in that: in the third step, the distance measurement formula is as follows:
RSSI(d)＝RSSI(d 0 )-10nlg d。
4. the RFID indoor positioning method based on BP and DNN dual neural networks as claimed in claim 1, characterized in that: in the fifth step, because it is very difficult to collect a large amount of data under actual conditions, a k-fold cross validation method is adopted for training, the collected data set D is divided into k mutually exclusive subsets with similar sizes, one subset is reserved as a test set, the other k-1 subsets are used as training sets, cross validation is repeated for k times, and the average value of k test results is returned.
CN201811255127.2A 2018-10-26 2018-10-26 RFID indoor positioning method based on BP and DNN double neural networks Active CN109444813B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811255127.2A CN109444813B (en) 2018-10-26 2018-10-26 RFID indoor positioning method based on BP and DNN double neural networks

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811255127.2A CN109444813B (en) 2018-10-26 2018-10-26 RFID indoor positioning method based on BP and DNN double neural networks

Publications (2)

Publication Number Publication Date
CN109444813A CN109444813A (en) 2019-03-08
CN109444813B true CN109444813B (en) 2022-10-18

Family

ID=65548691

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811255127.2A Active CN109444813B (en) 2018-10-26 2018-10-26 RFID indoor positioning method based on BP and DNN double neural networks

Country Status (1)

Country Link
CN (1) CN109444813B (en)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109996172A (en) * 2019-03-14 2019-07-09 北京工业大学 One kind being based on BP neural network precision indoor positioning system and localization method
CN110990500B (en) * 2019-03-29 2023-12-15 天维讯达(湖南)科技有限公司 Propagation path model map establishment method and path loss determination method
CN111741524B (en) * 2019-06-04 2021-08-31 腾讯科技（深圳）有限公司 Positioning method, positioning device, computer readable storage medium and computer equipment
CN110320513A (en) * 2019-07-05 2019-10-11 南京简睿捷软件开发有限公司 A kind of production factors positioning system and method for large area workshop based on RFID
CN110381580B (en) * 2019-07-25 2021-03-26 上海开域信息科技有限公司 WiFi positioning method based on proportion optimization
CN110909873B (en) * 2019-10-08 2022-07-22 北京建筑大学 Method and device for denoising passive RFID (radio frequency identification) mobile ranging data and computer equipment
CN110888105B (en) * 2019-10-09 2023-03-24 哈尔滨工程大学 DOA estimation method based on convolutional neural network and received signal strength
CN111031477A (en) * 2019-12-25 2020-04-17 中山大学 Millimeter wave indoor positioning method based on deep learning
CN111724495A (en) * 2020-04-30 2020-09-29 杭州任联科技有限公司 Non-inductive attendance system and method based on dual-orientation gateway, RFID and BP neural network
CN111523667B (en) * 2020-04-30 2023-06-27 天津大学 RFID positioning method based on neural network
US20230420838A1 (en) * 2020-11-13 2023-12-28 Telefonaktiebolaget Lm Ericsson (Publ) Ue driven antenna tilt
CN113203419B (en) * 2021-04-25 2023-11-10 重庆大学 Indoor inspection robot correction positioning method based on neural network
CN113640740B (en) * 2021-08-04 2023-09-29 成都诚骏科技有限公司 Indoor high-precision positioning method for intelligent warehouse management system
CN115113168A (en) * 2022-08-25 2022-09-27 南京宇安防务科技有限公司 Radar clutter suppression method based on neural network

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103945533B (en) * 2014-05-15 2016-08-31 济南嘉科电子技术有限公司 Wireless real time position localization methods based on big data
US9569589B1 (en) * 2015-02-06 2017-02-14 David Laborde System, medical item including RFID chip, data collection engine, server and method for capturing medical data
CN106959444B (en) * 2017-03-07 2019-04-26 上海工程技术大学 A kind of RFID indoor locating system and method based on artificial neural network
CN107247260B (en) * 2017-07-06 2019-12-03 合肥工业大学 A kind of RFID localization method based on adaptive depth confidence network

Also Published As

Publication number Publication date
CN109444813A (en) 2019-03-08

Similar Documents

Publication Publication Date Title
CN109444813B (en) RFID indoor positioning method based on BP and DNN double neural networks
CN108989976B (en) Fingerprint positioning method and system in intelligent classroom
CN109275095B (en) Bluetooth-based indoor positioning system, positioning equipment and positioning method
CN112533149B (en) Moving target positioning algorithm based on UWB mobile node
CN108871332A (en) A kind of RFID indoor locating system and method based on XGBoost
CN110996280B (en) RFID indoor positioning fingerprint database updating system and method
CN110536257B (en) Indoor positioning method based on depth adaptive network
CN104038901A (en) Indoor positioning method for reducing fingerprint data acquisition workload
CN111586605B (en) KNN indoor target positioning method based on adjacent weighted self-adaptive k value
CN110046834A (en) Workplace pernicious gas health risk quantitative evaluating system and its appraisal procedure
CN102967848B (en) Positioning method based on distance relationship library and received signal intensity
CN106932752A (en) A kind of real-time indoor orientation method based on RF phse monitoring
CN103415027B (en) WIFI indoor signal distribution model automatically selects and localization method
CN115824993B (en) Method and device for determining water body chemical oxygen demand, computer equipment and medium
CN105044659B (en) Indoor positioning device and method based on ambient light spectrum fingerprint
CN105866732A (en) Improved MK model and WKNN algorithm combined mixed indoor positioning method
CN109541537B (en) Universal indoor positioning method based on ranging
Liu et al. CTSLoc: An indoor localization method based on CNN by using time-series RSSI
CN109302674B (en) WiFi indoor positioning method based on multiple filtering
Ghozali et al. Indoor positioning system using regression-based fingerprint method
CN109766969B (en) RFID indoor positioning algorithm based on asynchronous dominant motion evaluation
CN116887396A (en) Training method of position prediction model, terminal positioning method and device
CN116234005A (en) Indoor positioning method for multiple buildings
CN114710831B (en) RFID label positioning system based on deep learning
CN104635206B (en) A kind of method and device of wireless location

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