CN112399366A - Indoor positioning method based on Hankel matrix and WKNN variance extraction - Google Patents

Indoor positioning method based on Hankel matrix and WKNN variance extraction Download PDF

Info

Publication number
CN112399366A
CN112399366A CN202010462500.2A CN202010462500A CN112399366A CN 112399366 A CN112399366 A CN 112399366A CN 202010462500 A CN202010462500 A CN 202010462500A CN 112399366 A CN112399366 A CN 112399366A
Authority
CN
China
Prior art keywords
matrix
variance
wknn
signal
noise
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
CN202010462500.2A
Other languages
Chinese (zh)
Inventor
郭正硕
吴锦州
潘甦
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 CN202010462500.2A priority Critical patent/CN112399366A/en
Publication of CN112399366A publication Critical patent/CN112399366A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/33—Services specially adapted for particular environments, situations or purposes for indoor environments, e.g. buildings
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/21—Design, administration or maintenance of databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • 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
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising

Abstract

The invention discloses an indoor positioning method based on Hankel matrix and WKNN variance extraction, and provides a noise reduction method divided into an off-line stage and an on-line stage, wherein Hankel matrix is used for reconstructing a fingerprint database to perform primary noise reduction in the off-line stage; and in the online stage, the influence caused by the instability of the AP point transmitting signal is removed by using the improved WKNN algorithm, so that the positioning precision is improved, and the defects of more noise and poor filtering treatment effect in the positioning process in the prior art are overcome.

Description

Indoor positioning method based on Hankel matrix and WKNN variance extraction
Technical Field
The invention relates to the field of indoor positioning, in particular to an indoor positioning method based on Hankel matrix and WKNN variance extraction.
Background
Currently, the positioning method based on the WLAN mainly includes a triangle algorithm and position fingerprint identification, where the triangle algorithm estimates the position of a target by using distance information between the target to be measured and at least three known APs (wireless access points), where the distance is estimated by measuring RSS from the APs. WLAN positioning based on the triangle algorithm relies heavily on accurate signal transmission loss models. The factors influencing signal transmission are many, the signal transmission loss models in different environments are quite different, and the establishment of an accurate loss model suitable for practical application is very difficult. Therefore, wireless positioning based on the triangle algorithm is more difficult in implementation.
When a fingerprint database is constructed for matrix recovery, several well-known matrix recovery algorithms can be implemented in noise reduction on the fingerprint database, including an Iteration Threshold (IT), an accelerated near-end gradient (APG), a Singular Value Threshold (SVT) and an imprecision-enhanced lagrangian multiplier (IALM). IT has a simple iterative form and low computational complexity. But its convergence speed is relatively slow and the iteration step of each iteration is uncertain. The APG is a first order algorithm of the Nesterov rule, and can convert an optimization model into an unconstrained form. It requires a complete singular value decomposition in each iteration, which is time consuming. SVT is mainly used to solve the matrix filling problem by rank minimization. However, it is often an NP challenge and does not achieve the desired noise reduction. IALM does not require an exact solution to the original problem during each iteration, which greatly reduces the number of singular value decompositions and saves computation time, but its noise reduction performance is not as good as expected.
When fingerprint matching is performed to determine the user position, in the conventional WKNN, a weight is given to each fingerprint according to the contribution degree of each sampling point to an unknown node, and the coordinates of the unknown node are estimated by the sum of products of the coordinates of the selected sampling point and the corresponding weight. The contribution degree is closely related to the Euclidean distance between the unknown node and the fingerprint record, and the smaller the Euclidean distance is, the larger the contribution degree is, and the larger the weight value is. However, when the similarity calculation is performed using the euclidean distance, the difference in the signal strength is not necessarily caused by the distance between the physical positions, but may be caused by the fluctuation of the signal strength itself. The sources of the fluctuations are roughly divided into two categories, one being the influence of the external environment of the positioning system: such as wall partition, personnel flow, same frequency interference and the like in indoor environment, and the internal influence of the positioning system: such as poor power stability of the transmitted signal of the AP. Document [2] indicates that the RSSI probability distribution presents a certain gaussian distribution characteristic, and document [1] proposes a WKNN indoor positioning method based on the RSSI distribution overlapping similarity on the basis of [2], that is, a similarity threshold is set to select a neighbor fingerprint point as a sampling point by using the relationship between the RSSI gaussian probability distribution overlapping similarity and the distance, so as to obtain a positioning result, thereby solving the problems that the traditional maximum similarity method may bring large errors and the positioning time is long, but actually the signal intensity distribution on the reference point is not a standard gaussian distribution and is also influenced by various interferences, so that the correlation relationship between the signal intensities of the APs also needs to be considered.
Disclosure of Invention
The purpose of the invention is as follows: the invention aims to provide an indoor positioning method based on Hankel matrix and WKNN variance extraction, and improve the indoor positioning accuracy.
The technical scheme is as follows: the invention discloses an indoor positioning method based on Hankel matrix and WKNN variance extraction, which comprises the following steps:
s1: decomposing and reconstructing singular values of a matrix by utilizing the structural specificity that a Hankel matrix has equal elements in each ascending diagonal line from left to right, filtering to reduce noise and constructing a fingerprint database at an offline stage;
s2: comparing the self-adaptive Kalman filtering with a fingerprint database constructed in an off-line stage, and filtering noise by using a Sage-Husa self-adaptive filtering algorithm;
s3: and (3) taking variance from the signal strength measured for multiple times in an off-line stage, calculating the full time, then improving an Euclidean distance formula by a weight coefficient, and eliminating the fluctuation of the AP by using an improved WKNN algorithm to obtain the position coordinate of the node to be positioned.
Preferably, the fingerprint database constructed by decomposing, reconstructing and filtering the matrix singular values in S1 includes the following steps:
s1.1, forming a Hankel matrix by RSS vectors of all reference points from the same AP, namely, each ascending diagonal line of the matrix from left to right has equal elements;
s1.2, according to the irrelevance between the real signal and the noise signal and the characteristics that the energy of the real signal is concentrated and the energy of the noise signal is dispersed, carrying out singular value decomposition on the Henkel matrix H to obtain:
Figure BDA0002511493100000021
in the formula of Um×mAnd Vn×nIs an orthogonal matrix; sigmam×nIn the form of a non-diagonal matrix,
Figure BDA0002511493100000022
S＝diag(σ1,σ2,σ3,...,σr)，σisingular values of matrix H;
s1.3, setting the threshold of singular value by using a characteristic mean value method, namely, selecting the first singular value as an estimated ideal signal space, and dividing H into the estimated ideal signal space
Figure BDA0002511493100000023
And estimated noise space
Figure BDA0002511493100000024
After the singular value decomposition of the Hankel matrix containing noise, the obtained singular value is set as sigmaiIs a square matrix AATCharacteristic value λ ofiThe square root of (c), i.e.:
Figure BDA0002511493100000025
s1.4. extracting x ═ S (k) ═ S1,s2,...,sn]The average value s of each sub diagonal in the series of sub diagonals is used to reconstruct the required signal sequence x ═ s (k) ═ s1,s2,...,sn]，x＝s(k)＝[s1,s2,...,sn]Can be expressed as follows, n is a matrix
Figure BDA0002511493100000026
The number of columns: x ═ s (k) ═ s1,s2,...,sn]. Setting the estimated signal vector
Figure BDA0002511493100000027
All will be
Figure BDA0002511493100000028
Obtaining a fingerprint database after noise reduction according to sequential arrangement:
Figure BDA0002511493100000029
preferably, the noise filtering in S2 by using the Sage-Husa adaptive filtering algorithm is to reprocess the primary filter matrix obtained in S1 and introduce the section update parameter dkAnd a forgetting factor b to improve the positioning accuracy of the filtering algorithm.
Preferably, the extracting the variance in S3 is to perform multiple sample acquisitions on each sampling point when the fingerprint library is established in the offline stage, where the signal strength received by each AP at each sampling point is a sample set, which includes the signal strength acquired multiple times at the point, and the variance may be calculated by using the sample set, and if the signal strength information set received by the tth AP at the sampling point j is:
Figure BDA0002511493100000031
wherein z is the amount of sample at each sampling point; the RSS vector of any sampling point j in the fingerprint library is:
Figure BDA0002511493100000032
wherein N is the number of APs, thereby obtaining the received signal strength rssjAnd the received signal variance σjThe vector of (a):
Figure BDA0002511493100000033
preferably, the weight calculation in S3 takes the reciprocal of the variance as a coefficient, and normalizes the coefficients of all APs at a sampling point j:
Figure BDA0002511493100000034
has good effectOptionally, the modified Euclidean distance formula in S3 is a weight coefficient
Figure BDA0002511493100000035
Improving an Euclidean distance calculation formula, wherein the weighted distance between a point i to be positioned and a sampling point j is as follows:
Figure BDA0002511493100000036
by using dijAnd screening out sample points meeting the conditions.
Has the advantages that: according to the method, the Hankel matrix is used for reconstructing the fingerprint database to perform primary noise reduction in the off-line stage, the influence caused by instability of AP point transmitting signals is removed by using the improved WKNN algorithm in the on-line stage, and the defects that in the prior art, the noise is more in the positioning process and the filtering processing effect is poor are overcome, so that the positioning precision is improved.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
As shown in fig. 1, the flow chart of the present invention includes the following steps:
s1: the method comprises the following steps of decomposing and reconstructing singular values of a matrix by utilizing the structural specificity that a Hankel matrix has equal elements in each ascending diagonal line from left to right, filtering to reduce noise, and constructing a fingerprint database in an offline stage, wherein the method specifically comprises the following steps:
s1.1, forming a Hankel matrix by RSS vectors of all reference points from the same AP, namely, each ascending diagonal line of the matrix from left to right has equal elements;
s1.2, according to the irrelevance between the real signal and the noise signal and the characteristics that the energy of the real signal is concentrated and the energy of the noise signal is dispersed, carrying out singular value decomposition on the Henkel matrix H to obtain:
Figure BDA0002511493100000037
in the formula of Um×mAnd Vn×nIs an orthogonal matrix; sigmam×nIn the form of a non-diagonal matrix,
Figure BDA0002511493100000041
S＝diag(σ1,σ2,σ3,...,σr)，σisingular values of matrix H;
s1.3, setting the threshold of singular value by using a characteristic mean value method, namely, selecting the first singular value as an estimated ideal signal space, and dividing H into the estimated ideal signal space
Figure BDA0002511493100000042
And estimated noise space
Figure BDA0002511493100000043
After the singular value decomposition of the Hankel matrix containing noise, the obtained singular value is set as sigmaiIs a square matrix AATCharacteristic value λ ofiThe square root of (c), i.e.:
Figure BDA0002511493100000044
s1.4. extracting x ═ S (k) ═ S1,s2,...,sn]The average value s of each sub diagonal in the series of sub diagonals is used to reconstruct the required signal sequence x ═ s (k) ═ s1,s2,...,sn]，x＝s(k)＝[s1,s2,...,sn]Can be expressed as follows, n is a matrix
Figure BDA0002511493100000045
The number of columns: x ═ s (k) ═ s1,s2,...,sn]. Setting the estimated signal vector
Figure BDA0002511493100000046
All will be
Figure BDA0002511493100000047
Obtaining a fingerprint database after noise reduction according to sequential arrangement:
Figure BDA0002511493100000048
s2: using adaptive kalman filterFiltering is compared with a fingerprint database constructed in an off-line stage, a primary filtering matrix obtained in S1 is reprocessed by using a Sage-Husa adaptive filtering algorithm, and a section updating parameter d is introducedkAnd a forgetting factor b to improve the positioning accuracy of the filtering algorithm.
S3: the method comprises the following steps of taking variance of signal strength measured for multiple times in an off-line stage, calculating full time, improving an Euclidean distance formula by a weight coefficient, eliminating fluctuation of an AP by using an improved WKNN algorithm, and obtaining a position coordinate of a node to be positioned, wherein the method specifically comprises the following steps:
s3.1, extracting variance, namely performing sample acquisition on each sampling point for multiple times when a fingerprint base is established in an off-line stage, wherein the signal intensity of each AP received by each sampling point is a sample set, the signal intensity of each AP received by each sampling point comprises the signal intensity acquired at the point for multiple times, and the variance can be calculated through the sample set, if the signal intensity information set of the tth AP received by the sampling point j is as follows:
Figure BDA0002511493100000049
wherein z is the amount of sample at each sampling point; the RSS vector of any sampling point j in the fingerprint database is as follows:
Figure BDA00025114931000000410
wherein N is the number of APs, thereby obtaining the received signal strength rssjAnd the received signal variance σjThe vector of (a):
Figure BDA00025114931000000411
s3.2, calculating the weight: taking the reciprocal of the variance as a coefficient, and carrying out normalization processing on the coefficients of all APs on the sampling point j:
Figure BDA00025114931000000412
s3.3 weight coefficient
Figure RE-GDA00026980358500000413
Improving an Euclidean distance calculation formula, wherein the weighted distance between a point i to be located and a sampling point j is as follows:
Figure RE-GDA00026980358500000414
by using dijAnd screening out sample points meeting the conditions, and finally obtaining the position coordinates of the nodes to be positioned by adopting a WKNN matching calculation formula.

Claims (6)

1. Indoor positioning method based on Hankel matrix and WKNN variance extraction, its characterized in that: the method comprises the following steps:
s1: decomposing and reconstructing matrix singular values by using the structural specificity that a Hankel matrix has equal elements in each ascending diagonal line from left to right, filtering to reduce noise, and constructing a fingerprint database at an off-line stage;
s2: comparing the self-adaptive Kalman filtering with a fingerprint database constructed in an off-line stage, and filtering noise by using a Sage-Husa self-adaptive filtering algorithm;
s3: and (3) taking variance from the signal strength measured for multiple times in an off-line stage, calculating the full time, then improving an Euclidean distance formula by a weight coefficient, and eliminating the fluctuation of the AP by using an improved WKNN algorithm to obtain the position coordinate of the node to be positioned.
2. The indoor positioning method based on Hankel matrix and WKNN variance extraction as claimed in claim 1, wherein: the fingerprint database constructed by decomposing, reconstructing and filtering the matrix singular values in S1 includes the following steps:
s1.1, forming a Hankel matrix by RSS vectors of all reference points from the same AP, namely each ascending diagonal line of the matrix from left to right has equal elements;
s1.2, according to the irrelevance between the real signal and the noise signal and the characteristics that the energy of the real signal is concentrated and the energy of the noise signal is dispersed, carrying out singular value decomposition on the Henkel matrix H to obtain:
Figure FDA0002511493090000011
in the formula of Um×mAnd Vn×nIs an orthogonal matrix; sigmam×nIs notThe diagonal matrix is a matrix of the angles of the,
Figure FDA0002511493090000012
S＝diag(σ1,σ2,σ3,...,σr)，σisingular values of matrix H;
s1.3, setting the threshold of singular value by using a characteristic mean value method, namely, selecting the first singular value as an estimated ideal signal space, and dividing H into the estimated ideal signal space
Figure FDA0002511493090000013
And estimated noise space
Figure FDA0002511493090000014
After the singular value decomposition of the Hankel matrix containing noise, the obtained singular value is set as sigmaiIs a square matrix AATCharacteristic value λ ofiThe square root of (c), i.e.:
Figure FDA0002511493090000015
s1.4. extracting x ═ S (k) ═ S1,s2,...,sn]The average value s of each sub diagonal in the series of sub diagonals is used to reconstruct the required signal sequence x ═ s (k) ═ s1,s2,...,sn]，x＝s(k)＝[s1,s2,...,sn]Can be expressed as follows, n is a matrix
Figure FDA0002511493090000016
The number of columns: x ═ s (k) ═ s1,s2,...,sn]. Setting the estimated signal vector
Figure FDA0002511493090000017
All will be
Figure FDA0002511493090000018
Obtaining a fingerprint database after noise reduction according to sequential arrangement:
Figure FDA0002511493090000019
3. the indoor positioning method based on Hankel matrix and WKNN variance extraction as claimed in claim 1, wherein: the step of filtering the noise by using the Sage-Husa adaptive filtering algorithm in the step S2 is to reprocess the primary filtering matrix obtained in the step S1 and introduce a section updating parameter dkAnd a forgetting factor b to improve the positioning accuracy of the filtering algorithm.
4. The indoor positioning method based on Hankel matrix and WKNN variance extraction as claimed in claim 1, wherein: the variance extraction in S3 is to perform multiple sample acquisitions on each sampling point when the fingerprint database is established in the offline stage, where the signal strength received by each AP by each sampling point is a sample set, which includes the signal strength acquired multiple times at the point, and the variance of the sample set can be calculated by using the sample set, where the signal strength information set received by the sampling point j by the tth AP is:
Figure FDA0002511493090000021
wherein z is the amount of sample at each sampling point; the RSS vector of any sampling point j in the fingerprint library is:
Figure FDA0002511493090000022
wherein N is the number of APs, thereby obtaining the received signal strength rssjAnd the received signal variance σjThe vector of (a):
Figure FDA0002511493090000023
5. the indoor positioning method based on Hankel matrix and WKNN variance extraction as claimed in claim 1, wherein: in the step S3, the calculated weight takes the reciprocal of the variance as a coefficient, and normalizes the coefficients of all APs at the sampling point j:
Figure FDA0002511493090000024
6. the indoor positioning method based on Hankel matrix and WKNN variance extraction as claimed in claim 1, wherein: the improved Euclidean distance formula in S3 is to use weight coefficient
Figure FDA0002511493090000025
Improving an Euclidean distance calculation formula, wherein the weighted distance between a point i to be located and a sampling point j is as follows:
Figure FDA0002511493090000026
by using dijAnd screening out sample points meeting the conditions.
CN202010462500.2A 2020-05-27 2020-05-27 Indoor positioning method based on Hankel matrix and WKNN variance extraction Pending CN112399366A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010462500.2A CN112399366A (en) 2020-05-27 2020-05-27 Indoor positioning method based on Hankel matrix and WKNN variance extraction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010462500.2A CN112399366A (en) 2020-05-27 2020-05-27 Indoor positioning method based on Hankel matrix and WKNN variance extraction

Publications (1)

Publication Number Publication Date
CN112399366A true CN112399366A (en) 2021-02-23

Family

ID=74603824

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010462500.2A Pending CN112399366A (en) 2020-05-27 2020-05-27 Indoor positioning method based on Hankel matrix and WKNN variance extraction

Country Status (1)

Country Link
CN (1) CN112399366A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113781559A (en) * 2021-08-31 2021-12-10 南京邮电大学 Robust abnormal matching point removing method and image indoor positioning method

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105933975A (en) * 2016-04-11 2016-09-07 南京邮电大学 WiFi fingerprint-based accuracy improved indoor positioning method
CN109916410A (en) * 2019-03-25 2019-06-21 南京理工大学 A kind of indoor orientation method based on improvement square root Unscented kalman filtering
CN110572875A (en) * 2019-09-16 2019-12-13 南京邮电大学 Wireless positioning method based on improved machine learning algorithm
CN110703205A (en) * 2019-10-14 2020-01-17 江苏帝一集团有限公司 Ultrashort baseline positioning method based on adaptive unscented Kalman filtering

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105933975A (en) * 2016-04-11 2016-09-07 南京邮电大学 WiFi fingerprint-based accuracy improved indoor positioning method
CN109916410A (en) * 2019-03-25 2019-06-21 南京理工大学 A kind of indoor orientation method based on improvement square root Unscented kalman filtering
CN110572875A (en) * 2019-09-16 2019-12-13 南京邮电大学 Wireless positioning method based on improved machine learning algorithm
CN110703205A (en) * 2019-10-14 2020-01-17 江苏帝一集团有限公司 Ultrashort baseline positioning method based on adaptive unscented Kalman filtering

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
张润轩: "基于WiFi的室内三维定位技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *
郭正硕: "基于机器学习的室内定位技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113781559A (en) * 2021-08-31 2021-12-10 南京邮电大学 Robust abnormal matching point removing method and image indoor positioning method
CN113781559B (en) * 2021-08-31 2023-10-13 南京邮电大学 Robust abnormal matching point eliminating method and image indoor positioning method

Similar Documents

Publication Publication Date Title
CN107948930B (en) Indoor positioning optimization method based on position fingerprint algorithm
CN108696331B (en) Signal reconstruction method based on generation countermeasure network
CN108470089B (en) Complex signal time delay estimation method based on least square sample fitting
CN107817465A (en) The DOA estimation method based on mesh free compressed sensing under super-Gaussian noise background
CN108520310B (en) Wind speed forecasting method of G-L mixed noise characteristic v-support vector regression machine
CN112036239B (en) Radar signal working mode identification method and system based on deep learning network
CN109855875B (en) Rolling bearing operation reliability prediction method
CN111983927A (en) Novel maximum entropy ellipsoid collective filtering method
CN114449452A (en) Indoor positioning algorithm for heterogeneous equipment
CN108921170B (en) Effective image noise detection and denoising method and system
CN112861066A (en) Machine learning and FFT (fast Fourier transform) -based blind source separation information source number parallel estimation method
CN113504505B (en) One-dimensional DOA estimation method suitable for low signal-to-noise ratio environment
CN112399366A (en) Indoor positioning method based on Hankel matrix and WKNN variance extraction
CN114757224A (en) Specific radiation source identification method based on continuous learning and combined feature extraction
CN110045363B (en) Multi-radar track association method based on relative entropy
CN111175692B (en) Discrete sparse Bayesian DOA estimation method based on layered synthesis Lasso prior model
CN114584230B (en) Predictive channel modeling method based on countermeasure network and long-term and short-term memory network
CN108834043B (en) Priori knowledge-based compressed sensing multi-target passive positioning method
CN116559579A (en) Improved VMD and Teager energy operator fault positioning method
CN111859241B (en) Unsupervised sound source orientation method based on sound transfer function learning
CN111160464B (en) Industrial high-order dynamic process soft measurement method based on multi-hidden-layer weighted dynamic model
CN111666688B (en) Corrected channel estimation algorithm combining angle mismatch with sparse Bayesian learning
CN113406560A (en) Angle and frequency parameter estimation method of incoherent distributed broadband source
CN111083632A (en) Ultra-wideband indoor positioning method based on support vector machine
CN112954637A (en) Target positioning method under condition of uncertain anchor node position

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

Application publication date: 20210223

RJ01 Rejection of invention patent application after publication