CN111918388A - CSI fingerprint passive positioning method based on depth separable convolution - Google Patents

CSI fingerprint passive positioning method based on depth separable convolution Download PDF

Info

Publication number
CN111918388A
CN111918388A CN202010826302.XA CN202010826302A CN111918388A CN 111918388 A CN111918388 A CN 111918388A CN 202010826302 A CN202010826302 A CN 202010826302A CN 111918388 A CN111918388 A CN 111918388A
Authority
CN
China
Prior art keywords
csi
convolution
depth
target position
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
CN202010826302.XA
Other languages
Chinese (zh)
Inventor
孙力娟
荀文婧
韩崇
郭剑
肖甫
王娟
周剑
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 CN202010826302.XA priority Critical patent/CN111918388A/en
Publication of CN111918388A publication Critical patent/CN111918388A/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
    • H04W64/006—Locating users or terminals or network equipment for network management purposes, e.g. mobility management with additional information processing, e.g. for direction or speed determination
    • 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
    • G01S5/0252—Radio frequency fingerprinting
    • 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
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/309—Measuring or estimating channel quality parameters

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Quality & Reliability (AREA)
  • Electromagnetism (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The CSI fingerprint passive positioning method based on the depth separable convolution is realized by taking a CSI data construction characteristic image as a position fingerprint and identifying different positions through a convolution neural network. The method comprises the following steps: and in an off-line training stage, the amplitude of CSI in the MIMO system is extracted, then a CSI characteristic graph similar to an RGB three-channel image is constructed to serve as position fingerprints of different positions, and then a convolutional neural network is designed by utilizing deep separable convolution to learn the CSI characteristics of the different positions. And in the on-line positioning stage, CSI data of a target position are collected to construct a target position characteristic image, and then the position of the target is predicted by using a trained convolutional neural network. The method is based on the position fingerprints fusing the CSI time domain, the frequency domain and the space domain, and adopts the depth separable convolution, so that higher position identification rate and lower positioning time delay can be obtained.

Description

CSI fingerprint passive positioning method based on depth separable convolution
Technical Field
The invention relates to the technical field of indoor positioning, in particular to a CSI fingerprint passive positioning method based on depth separable convolution.
Background
Mobile smart devices and wireless networks have penetrated aspects of human productive life, and Location Based Services (LBS) are gradually becoming indispensable services for human life. Although the Global Positioning System (GPS) has been unique in the field of outdoor navigation and Positioning, the severe fading of the GPS signal caused by the reinforced cement makes it difficult to use in indoor navigation and Positioning. Indoor positioning has important research significance and practical value, attracts the research enthusiasm of a large number of researchers at home and abroad, and simultaneously shows a large number of indoor positioning schemes, and the computer vision technology, the infrared technology, the ultrasonic technology, the ultra-wideband technology, the Bluetooth, the Zigbee, the RFID and the like are mainly applied to indoor positioning at present. However, most of these positioning schemes require a large amount of hardware devices, which limits the universality of the positioning schemes. With the wide spread of Wi-Fi infrastructure and wireless devices, Wi-Fi based indoor positioning schemes are becoming mainstream.
Received Signal Strength (RSS) is taken as a main energy characteristic measured value of a wireless Signal, can be directly obtained from a large number of wireless terminals, and is widely applied to Wi-Fi-based indoor positioning systems. In an indoor environment, however, wireless signals travel through multiple paths between a transmitter and a receiver due to the presence of obstacles such as walls, furniture, and the like. The signal received by the receiver is the effect of the superposition of the multiple path signals. RSS, which is an average value of multipath signals, is likely to fluctuate in an indoor environment and is poor in stability. This severely limits the reliability and positioning accuracy of RSS based indoor positioning schemes. Channel State Information (CSI) is a finer granularity of physical layer Information than RSS. Each set of CSI characterizes the amplitude and phase of an Orthogonal Frequency Division Multiplexing (OFDM) subcarrier, with better static stability and dynamic sensitivity.
In fingerprint positioning, there is generally a certain difference between location fingerprints at different physical locations. Thus, fingerprint localization can be viewed simply as a sort task. The kNN and Bayesian classification are used as mainstream classification algorithms and are widely applied to fingerprint positioning. However, these conventional classification algorithms have certain disadvantages. kNN as the data dimension increases, the computational overhead can become significant. For the Bayes classification, a conditional independence hypothesis premise is needed, but the situation is not always the case, which seriously restricts the accuracy of the Bayes classification.
In recent years, with the development of deep learning, more and more researchers apply neural networks to fingerprint localization. Compared with the traditional classification method, the classification method based on deep learning tends to have higher accuracy. However, most of the existing CSI fingerprint positioning schemes based on deep learning need to spend a lot of time in the off-line training stage of the model, and the on-line positioning delay is high, so that the requirement of people on the real-time performance of the positioning service cannot be met.
Disclosure of Invention
The invention aims to provide a CSI fingerprint passive positioning method based on depth separable convolution. The invention has lower positioning time delay and higher positioning precision.
The CSI fingerprint passive positioning method based on the depth separable convolution comprises the following steps:
step 1: dividing a scene area, selecting a reference point, and acquiring reference point coordinates and CSI data through a transmitter and a receiver;
step 2: setting transmitting and receiving antenna pairs of a transmitter and a receiver, and constructing CSI characteristic images for CSI data of each group of transmitting and receiving antenna pairs;
and step 3: designing a convolutional neural network based on depth separable convolution, training the neural network by using a CSI (channel state information) feature image of a reference point, and optimizing by adopting an Adam algorithm during training;
and 4, step 4: collecting target position CSI data, and constructing a CSI characteristic image of a target position according to the step 2;
and 5: and (3) inputting the CSI characteristic image of the target position into the neural network trained in the step (3), taking the probability of the target position at each reference point as a weight by the neural network, and calculating the weighted average of the coordinates of each reference point to obtain the final predicted target position coordinate.
Further, in step 2, using MIMO information, for data of each group of transmit-receive antenna pairs, a sliding window with a size of 30 is used to select 30 consecutive CSI data packets, to extract amplitudes of 30 subcarriers in each data packet, and to calculate an amplitude difference between each subcarrier and a previous subcarrier, and an amplitude difference between a first subcarrier and a last subcarrier. Constructing the amplitude difference of the CSI subcarriers into a 30 multiplied by 3 three-channel CSI characteristic image, wherein the data of each group of antenna pairs corresponds to one channel of the CSI characteristic image;
amplitude of sub-carrier i of jth data packet is
Figure BDA0002636307880000031
The amplitude difference of adjacent subcarriers can be expressed as:
Figure BDA0002636307880000032
one feature image of the CSI may be represented as:
Figure BDA0002636307880000041
where N is the number of subcarriers.
Further, in step 3, the first layer of the network structure adopts standard convolution, and the convolution kernel size is 3 × 3. Then, 3 depth-separable convolutions are employed, each depth-separable convolution comprising a depth convolution and a point-by-point convolution. The step size of the standard convolution, the point-by-point convolution and the first depth convolution is set to 1 and the depth convolution kernel size is 3 x 3. The second and third depth convolution step sizes are set to 2. Each depth convolution and point-by-point convolution is followed by a normalization and ReLU operation. The last of the network is a pooling layer and a full-link layer, the output size of which is N, where N is the number of reference points.
Further, in step 5, assuming that there are N reference points in total, the position of the ith reference point is LiThe probability of the target position at the ith reference point is PiThen predicted target position
Figure BDA0002636307880000043
Can be expressed as:
Figure BDA0002636307880000042
the invention achieves the following beneficial effects: a CSI fingerprint passive positioning method based on depth separable convolution is provided, the MIMO information of Wi-Fi is utilized, the amplitude difference of CSI is extracted to construct a CSI characteristic image to serve as a position fingerprint of each position, a convolution neural network is designed based on the depth separable convolution to realize positioning, and the method has low positioning time delay and high positioning accuracy.
Drawings
Fig. 1 is a schematic diagram of a CSI fingerprint passive location method based on depth separable convolution according to an embodiment of the present invention.
Fig. 2 is an experimental scenario layout diagram in an embodiment of the present invention.
Fig. 3 is a CSI characteristic diagram based on subcarrier amplitude differences in the embodiment of the present invention.
Detailed Description
The technical scheme of the invention is further explained in detail by combining the drawings in the specification.
The invention provides a CSI fingerprint passive positioning method based on depth separable convolution, which comprises the following steps:
step 1: as shown in fig. 2, an indoor scene area is divided, a reference point is selected, and reference point coordinates and CSI data are collected.
Step 2: and constructing a CSI characteristic image by using CSI data of different antenna pairs of a transmitter and a receiver. For data at each position, 3000 data packets are collected, a plurality of groups of continuous data packets in time are selected by using a sliding window with the size of 30, and CSI subcarrier amplitude in each data packet is extracted. The amplitude of the ith CSI subcarrier of packet j is
Figure BDA0002636307880000051
Calculating amplitude difference of adjacent subcarriers
Figure BDA0002636307880000052
And constructs CSI feature images of different locations as shown in fig. 3.
Specifically, in a smoothly fading narrowband channel, the OFDM system can be represented as:
yi＝Hxi+Ni
wherein y isiAnd xiRespectively representing signal vectors at the receiver end and the transmitter end, H being a channel state information matrix, NiRepresenting a noise vector.
For estimating the channel state information matrix, the transmitter sends a known pilot sequence x1，x2，...，xnThe combined received signal vector Y may be expressed as:
Y＝[y1,y2,...,yn]＝HX+N
the channel state information matrix H can therefore be approximated as:
Figure BDA0002636307880000061
for an OFDM system with N subcarriers, H may be expressed as:
H＝[H1,H2,...,HN]
wherein
Figure BDA0002636307880000062
For MIMO systems with m transmit antennas, n receive antennas, HiIs a matrix of m × n dimensions, which can be expressed as:
Figure BDA0002636307880000063
wherein h ispq(p∈[1,m],q∈[1,n]) Corresponding to the complex number of subcarrier amplitudes and phases on the transmit antenna p and receive antenna q antenna streams. Thus, the CSI matrix is an m × N matrix.
Amplitude of sub-carrier i of jth data packet is
Figure BDA0002636307880000064
Figure BDA0002636307880000065
One feature image F of the CSI may be represented as:
Figure BDA0002636307880000066
where N is the number of subcarriers 30.
And step 3: convolutional neural networks are designed based on deep separable convolutions. And training the neural network by using the CSI characteristic image of the reference point. And during network training, optimizing by adopting an Adam algorithm.
The first layer of the convolutional neural network uses standard convolution with a convolution kernel size of 3 x 3. Then, 3 depth-separable convolutions are employed, each depth-separable convolution comprising a depth convolution and a point-by-point convolution. The step size of the standard convolution, the point-by-point convolution and the first depth convolution is set to 1 and the depth convolution kernel size is 3 x 3. The second and third depth convolution step sizes are set to 2. Each depth convolution and point-by-point convolution is followed by a normalization and ReLU operation. The last of the network is a pooling layer and a full-link layer, and the output size of the full-link layer is the number of the reference points. The network parameters are shown in the following table:
Figure BDA0002636307880000071
and 4, step 4: collecting target position CSI data, selecting 30 data packets continuous in time according to step 2, extracting CSI subcarrier amplitude in each data packet, and calculating amplitude difference of adjacent subcarriers
Figure BDA0002636307880000072
And constructs a CSI feature image of the target location. And constructing a CSI characteristic image of the target position.
And 5: and (3) inputting the CSI characteristic image of the target position into the neural network trained in the step (3). The neural network will output the probability of the target location at each reference point. And taking the probability of the target position at each reference point as weight, and calculating the weighted average of the coordinates of each reference point to obtain the final predicted target position coordinate. Assuming that there are N reference points in total, the position of the ith reference point is LiThe probability of the target position at the ith reference point is PiThen predicted target position
Figure BDA0002636307880000082
Can be expressed as:
Figure BDA0002636307880000081
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. The CSI fingerprint passive positioning method based on the depth separable convolution is characterized in that: the positioning method comprises the following steps:
step 1: dividing a scene area, selecting a reference point, and acquiring reference point coordinates and CSI data through a transmitter and a receiver;
step 2: setting transmitting and receiving antenna pairs of a transmitter and a receiver, and constructing CSI characteristic images for CSI data of each group of transmitting and receiving antenna pairs;
and step 3: designing a convolutional neural network based on depth separable convolution, training the neural network by using a CSI (channel state information) feature image of a reference point, and optimizing by adopting an Adam algorithm during training;
and 4, step 4: collecting target position CSI data, and constructing a CSI characteristic image of a target position according to the step 2;
and 5: and (3) inputting the CSI characteristic image of the target position into the neural network trained in the step (3), taking the probability of the target position at each reference point as a weight by the neural network, and calculating the weighted average of the coordinates of each reference point to obtain the final predicted target position coordinate.
2. The method of claim 1, wherein the depth separable convolution-based CSI fingerprint passive location method comprises: in step 2, using MIMO information, for data of each group of transmit-receive antenna pairs, selecting continuous 30 CSI data packets by using a sliding window with a size of 30, extracting amplitudes of 30 subcarriers in each data packet, and calculating an amplitude difference between each subcarrier and a previous subcarrier, and an amplitude difference between a first subcarrier and a last subcarrier. Constructing the amplitude difference of the CSI subcarriers into a 30 multiplied by 3 three-channel CSI characteristic image, wherein the data of each group of antenna pairs corresponds to one channel of the CSI characteristic image;
amplitude of sub-carrier i of jth data packet is
Figure FDA0002636307870000011
The amplitude difference of adjacent subcarriers can be expressed as:
Figure FDA0002636307870000021
one feature image of the CSI may be represented as:
Figure FDA0002636307870000022
where N is the number of subcarriers.
3. The method of claim 1, wherein the depth separable convolution-based CSI fingerprint passive location method comprises: in step 3, the first layer of the network structure adopts standard convolution, and the convolution kernel size is 3 x 3. Then, 3 depth-separable convolutions are employed, each depth-separable convolution comprising a depth convolution and a point-by-point convolution. The step size of the standard convolution, the point-by-point convolution and the first depth convolution is set to 1 and the depth convolution kernel size is 3 x 3. The second and third depth convolution step sizes are set to 2. Each depth convolution and point-by-point convolution is followed by a normalization and ReLU operation. The last of the network is a pooling layer and a full-link layer, the output size of which is N, where N is the number of reference points.
4. The method of claim 1, wherein the depth separable convolution-based CSI fingerprint passive location method comprises: in step 5, assume that there are N reference points, and the position of the ith reference point is LiThe probability of the target position at the ith reference point is PiThen predicted target position
Figure FDA0002636307870000023
Can be expressed as:
Figure FDA0002636307870000024
CN202010826302.XA 2020-08-17 2020-08-17 CSI fingerprint passive positioning method based on depth separable convolution Pending CN111918388A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010826302.XA CN111918388A (en) 2020-08-17 2020-08-17 CSI fingerprint passive positioning method based on depth separable convolution

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010826302.XA CN111918388A (en) 2020-08-17 2020-08-17 CSI fingerprint passive positioning method based on depth separable convolution

Publications (1)

Publication Number Publication Date
CN111918388A true CN111918388A (en) 2020-11-10

Family

ID=73279646

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010826302.XA Pending CN111918388A (en) 2020-08-17 2020-08-17 CSI fingerprint passive positioning method based on depth separable convolution

Country Status (1)

Country Link
CN (1) CN111918388A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112689237A (en) * 2021-03-15 2021-04-20 四川北控聚慧物联网科技有限公司 Indoor positioning method based on WiFi
CN113822351A (en) * 2021-09-14 2021-12-21 南京邮电大学 Double-current convolution neural network target positioning method based on CSI image feature extraction
CN113939016A (en) * 2021-12-21 2022-01-14 广州优刻谷科技有限公司 Intelligent terminal indoor positioning method and system based on WIFI dual-frequency fusion
CN115166636A (en) * 2022-09-05 2022-10-11 南京信息工程大学 Factory personnel supervision method based on multi-feature channel state information edge calculation
CN115546710A (en) * 2022-08-09 2022-12-30 国网湖北省电力有限公司黄龙滩水力发电厂 Method, device and equipment for locating personnel in hydraulic power plant and readable storage medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107832834A (en) * 2017-11-13 2018-03-23 合肥工业大学 A kind of construction method of the WIFI indoor positioning fingerprint bases based on generation confrontation network
CN108650628A (en) * 2018-03-24 2018-10-12 西安电子科技大学 The indoor orientation method that ranging and fingerprint based on Wi-Fi network are combined
US20190059070A1 (en) * 2017-01-18 2019-02-21 Shenzhen University Indoor positioning method and system based on wireless receiver and camera
US20190362237A1 (en) * 2019-03-15 2019-11-28 Intel Corporation Machine learning techniques for precise position determination
CN111212379A (en) * 2020-01-06 2020-05-29 天津工业大学 Novel CSI indoor positioning method based on convolutional neural network
CN111479231A (en) * 2020-04-17 2020-07-31 西安交通大学 Indoor fingerprint positioning method for millimeter wave large-scale MIMO system

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190059070A1 (en) * 2017-01-18 2019-02-21 Shenzhen University Indoor positioning method and system based on wireless receiver and camera
CN107832834A (en) * 2017-11-13 2018-03-23 合肥工业大学 A kind of construction method of the WIFI indoor positioning fingerprint bases based on generation confrontation network
CN108650628A (en) * 2018-03-24 2018-10-12 西安电子科技大学 The indoor orientation method that ranging and fingerprint based on Wi-Fi network are combined
US20190362237A1 (en) * 2019-03-15 2019-11-28 Intel Corporation Machine learning techniques for precise position determination
CN111212379A (en) * 2020-01-06 2020-05-29 天津工业大学 Novel CSI indoor positioning method based on convolutional neural network
CN111479231A (en) * 2020-04-17 2020-07-31 西安交通大学 Indoor fingerprint positioning method for millimeter wave large-scale MIMO system

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
WENJING XUN等: "Depthwise Separable Convolution based Passive Indoor Localization using CSI Fingerprint", 《2020 IEEE WIRELESS COMMUNICATIONS AND NETWORKING CONFERENCE (WCNC)》 *
朱海等: "基于信道状态信息的WiFi环境感知技术", 《南京邮电大学学报(自然科学版)》 *
沙朝恒等: "一种基于矩阵补全的室内指纹定位算法", 《计算机科学》 *
郑瑞环等: "基于无线设备指纹的伪AP检测技术研究", 《信息安全研究》 *
鞠聪等: "基于深度可分离卷积结构的人脸表情识别研究", 《电脑与电信》 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112689237A (en) * 2021-03-15 2021-04-20 四川北控聚慧物联网科技有限公司 Indoor positioning method based on WiFi
CN113822351A (en) * 2021-09-14 2021-12-21 南京邮电大学 Double-current convolution neural network target positioning method based on CSI image feature extraction
CN113822351B (en) * 2021-09-14 2023-10-24 南京邮电大学 Dual-flow convolutional neural network target positioning method based on CSI image feature extraction
CN113939016A (en) * 2021-12-21 2022-01-14 广州优刻谷科技有限公司 Intelligent terminal indoor positioning method and system based on WIFI dual-frequency fusion
CN115546710A (en) * 2022-08-09 2022-12-30 国网湖北省电力有限公司黄龙滩水力发电厂 Method, device and equipment for locating personnel in hydraulic power plant and readable storage medium
CN115166636A (en) * 2022-09-05 2022-10-11 南京信息工程大学 Factory personnel supervision method based on multi-feature channel state information edge calculation
CN115166636B (en) * 2022-09-05 2022-12-20 南京信息工程大学 Factory personnel supervision method based on multi-feature channel state information edge calculation

Similar Documents

Publication Publication Date Title
CN111918388A (en) CSI fingerprint passive positioning method based on depth separable convolution
Hu et al. Experimental Analysis on Weight ${K} $-nearest neighbor indoor fingerprint positioning
Abdel-Nasser et al. MonoPHY: Mono-stream-based device-free WLAN localization via physical layer information
CN111479231B (en) Indoor fingerprint positioning method for millimeter wave large-scale MIMO system
CN105828289B (en) A kind of passive indoor orientation method based on channel state information
CN108594170B (en) WIFI indoor positioning method based on convolutional neural network identification technology
CN107769828A (en) A kind of the CSI MIMO indoor orientation methods and system of the extraction of feature based value
Li et al. Indoor localization based on CSI fingerprint by siamese convolution neural network
CN107727095A (en) 3D indoor orientation methods based on spectral clustering and weighting reverse transmittance nerve network
Li et al. Smartphone-based indoor localization with integrated fingerprint signal
CN113207089A (en) Position fingerprint positioning method based on CSI and crowdsourcing migration self-calibration updating
CN109640262B (en) Positioning method, system, equipment and storage medium based on mixed fingerprints
Rao et al. MSDFL: a robust minimal hardware low-cost device-free WLAN localization system
CN108519578A (en) A kind of indoor positioning fingerprint base construction method based on intelligent perception
CN113852922A (en) High-precision indoor positioning method for WiFi signal direct line-of-sight propagation path excavation
Ishida et al. Proposal of separate channel fingerprinting using Bluetooth low energy
Shih et al. Intelligent radio map management for future WLAN indoor location fingerprinting
Bahadori et al. ReWiS: Reliable Wi-Fi sensing through few-shot multi-antenna multi-receiver CSI learning
Wei et al. Enriching indoor localization fingerprint using a single AC-GAN
Fan et al. Single-site indoor fingerprint localization based on MIMO-CSI
CN112867021B (en) Improved TrAdaBoost-based indoor positioning method for transfer learning
Yeung et al. Enhanced fingerprint-based location estimation system in wireless LAN environment
CN111523571B (en) Semi-supervised learning positioning method for distributed large-scale multi-antenna system
Deng et al. Joint access point selection and local discriminant embedding for energy efficient and accurate Wi-Fi positioning
Xu et al. Variance-based fingerprint distance adjustment algorithm for indoor localization

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: 20201110