CN111199162A - RFID reader fault self-adaptive positioning method - Google Patents
RFID reader fault self-adaptive positioning method Download PDFInfo
- Publication number
- CN111199162A CN111199162A CN202010028807.1A CN202010028807A CN111199162A CN 111199162 A CN111199162 A CN 111199162A CN 202010028807 A CN202010028807 A CN 202010028807A CN 111199162 A CN111199162 A CN 111199162A
- Authority
- CN
- China
- Prior art keywords
- fuzzy
- output
- rfid reader
- online
- fault
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K7/00—Methods or arrangements for sensing record carriers, e.g. for reading patterns
- G06K7/10—Methods or arrangements for sensing record carriers, e.g. for reading patterns by electromagnetic radiation, e.g. optical sensing; by corpuscular radiation
- G06K7/10009—Methods or arrangements for sensing record carriers, e.g. for reading patterns by electromagnetic radiation, e.g. optical sensing; by corpuscular radiation sensing by radiation using wavelengths larger than 0.1 mm, e.g. radio-waves or microwaves
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Toxicology (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Electromagnetism (AREA)
- General Health & Medical Sciences (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a fault self-adaptive positioning method for an RFID reader. The method comprises the following steps: constructing a fault self-adaptive positioning system of the RFID reader; providing an online sequence fuzzy width learning system according to the RFID reader fault self-adaptive positioning system; an RFID reader fault self-adaption strategy is provided based on an online sequence fuzzy width learning system; and completing the RFID reader fault self-adaptive positioning based on the RFID reader fault self-adaptive strategy. By the aid of the RFID reader fault self-adaptive positioning method, data streams which continuously arrive in the environment can be processed, and subsequent data streams can be processed when partial readers are in fault, so that the aim of long-time stable positioning is achieved.
Description
Technical Field
The invention relates to the field of RFID, machine learning and positioning algorithms, in particular to a fault self-adaptive positioning method for an RFID reader.
Background
With the development of the internet of things technology, the demand of people for the application of the internet of things is rapidly increased, and among the technologies, the wireless positioning technology shows great liveliness in both military and civil aspects, and the wireless positioning technology and the wireless positioning service-based wireless positioning service play an increasingly greater role in the life of people. Among outdoor positioning technologies, the global positioning system is the most famous and representative positioning technology, and is widely used in military and civil applications. The demand of people on indoor positioning application is getting larger and larger, and the indoor long-time positioning demand has great potential.
Since the RFID has the characteristics of non-line-of-sight, non-contact, and capability of quickly identifying an object, it has certain advantages in indoor positioning. The basic RFID system consists of a label, an antenna and a reader, and the basic working principle is as follows: the RFID reader generates a magnetic field through an antenna, the tag enters the magnetic field, receives a radio frequency signal sent by the reader, and sends product information (a passive tag or a passive tag) stored in a chip or actively sends a radio frequency signal (an active tag or an active tag) of a certain frequency by means of energy obtained by electromagnetic induction.
At present, RFID positioning methods are mainly divided into two types, namely distance-based positioning and scene-based positioning. There are many distance-based RFID positioning methods, such as received signal strength indication, angle of arrival, time of arrival, and time difference of arrival, but these methods have high sensitivity to the environment and low accuracy (Ni L M, Zhang D, Souryal M r. RFID-based analysis and tracking technologies [ J ]. IEEE Wireless Communications,2011,18(2): 45-51.). Meanwhile, another type of RFID positioning method based on scene analysis is favored by researchers due to its advantages of high environmental adaptability, relatively low cost, etc. The LANDMARC is a classic RFID positioning algorithm (Ni LM, Liu Y, Lau Y C, et al LANDMARC: inductor location sensing active RFID [ C ]// Proceedings of the First IEEE International Conference on Perdynamic computing and Communications,2003 (PerCom 2003). IEEE,2003:407 + 415). The algorithm introduces the concept of a reference tag, and the partial tags with high signal strength indicating value similarity are listed as alternative reference tags, so as to obtain the position of a target tag through calculation. In the method based on scene analysis, some students use artificial neural networks and other methods to perform RFID positioning, and a passive RFID indoor positioning scheme (Kung H Y, Chaisits, wavelength N T M. optimization of an RFID positioning identification scheme [ J ]. International Journal of Communication Systems,2015,28(4): minus 644.) is provided, which combines the LANDMAC scheme and the BP neural network, and after LANDMAC positioning, the positioning result is further processed by using the BP neural network to obtain a more accurate positioning result. In addition, some researchers use fuzzy Neural Networks for RFID indoor localization, and use fuzzy Neural Networks to analyze The relation between The actual coordinates of The reference tags and The environmental errors, thereby adjusting The environmental parameters in The RFID localization system (Huang Y J, Chen C Y, Hong B W, et al. fuzzy Neural network based RFID index localization sensing technical [ C ]// The 2010International journal conference on Neural Networks (IJCNN). IEEE,2010: 1-5.).
However, in an actual RFID indoor positioning scenario, the positioning system needs to work for a long time, and therefore, a fault may occur. In order to solve the problem of RFID tag failure, some researchers have proposed a solution to handle independent permanent tag failure and regional permanent tag failure, but cannot handle RFID reader failure (Zhu W, Cao J, Xu Y, et. Fault-tolerant RFID reader localization based on passive RFID tags [ J ]. IEEETransactions on Parallel and Distributed Systems,2013,25(8): 2065) 2076.). When the RFID reader fails to be maintained in time, the existing positioning algorithm cannot maintain the precision of the original positioning system, so that the existing method needs to be improved, and a positioning method suitable for the RFID reader failure is provided to achieve the aim of long-time stable positioning.
Disclosure of Invention
The invention provides a self-adaptive fault positioning method for an RFID reader. The basic idea of the invention is to provide an online sequential fuzzy width learning system, which is improved to have online sequential learning capability, provide a pseudo-inverse update formula of an update matrix of a model and an update formula of a parameter matrix, and can process data streams which continuously arrive in an environment; and meanwhile, a fault self-adaption strategy of the RFID reader is provided, when part of the readers have faults, a conversion matrix is provided to process the coefficients and membership function centers randomly generated in the initialized fuzzy subsystem, new coefficient and membership function centers are generated, and the subsequent data streams are processed.
The purpose of the invention is realized by at least one of the following technical solutions.
An RFID reader fault self-adaptive positioning method comprises the following steps:
s1, constructing a fault self-adaptive positioning system of the RFID reader;
s2, providing an online sequence fuzzy width learning system according to the RFID reader fault self-adaptive positioning system;
s3, providing a fault self-adaptive strategy of the RFID reader based on the online sequence fuzzy width learning system;
s4, completing RFID reader fault adaptive positioning based on the RFID reader fault adaptive strategy.
Further, in step S1, the RFID reader fault adaptive positioning system includes M RFID readers, Q RFID reference tags, and a target tag to be positioned; the reference tags with known coordinate positions are deployed in a plane where the target tags to be positioned are located in an equilateral triangle mode, and the readers are deployed on four edges of the plane.
Further, in step S2, the online sequential fuzzy width learning system includes an input layer, a fuzzy subsystem, an enhancement layer and an output layer, wherein the fuzzy subsystem is a TS fuzzy model, and in the ith fuzzy subsystem, the weight is set Wherein xsMFor input data, s1, 2, N is the number of input data, M is the input data dimension, KiIn order to be able to blur the number of rules,for randomly generated coefficients, the membership function is a Gaussian functionWhereinRespectively, the center and width of the membership function, and the weighted output of the fuzzy rule isWhereinThe enhancement layer is a nonlinear transformation layer;
the number of the fuzzy subsystems is n, the number of the enhanced node groups is m, wherein the fuzzy rule number of the ith fuzzy subsystem is KiThe number of the j-th group of enhanced nodes is LjOf 1 atThe signal strength of the RFID reader reading the reference label at the s time is Generation of N0The bar input data isThe output data isWhereinIs the Nth0An array of signal strengths of the strip input data,is the Nth0Coordinates of the input data are obtained, and C is a characteristic dimension of the output data;
will N0The bar input data is used as the input of a fuzzy subsystem in the online sequential fuzzy width learning system, the fuzzy subsystem intermediate output and the fuzzy subsystem output are obtained through calculation, the fuzzy subsystem intermediate output is used as the input of an enhancement node, the enhancement layer output is further obtained through calculation, the fuzzy subsystem output and the enhancement layer output are connected with the output layer, and the weight to be calculated is obtained through pseudo-inverse.
Further, in step S3, the RFID reader fault adaptive policy is as follows:
when some RFID readers are in fault, the number of the RFID readers is changed into M' and 0<M' is less than or equal to M, then subsequently generated NaThe bar input data isRSSIs=(RSSIs1,RSSIs2,...,RSSIsM′),s=1,2,...,NaThe output data isWhereinIs the NthaSignal Strength array, RSSI, of the strip input datasM′For the signal strength of the mth reader reading the reference tag at the s-th time,is the NthaCoordinates of the bar input data;
after partial RFID readers are in fault, the characteristic dimension of input data is changed from M to M', and a conversion matrix T needs to be introduced into an initialization fuzzy subsystemCoefficients generated randomly by a functionAnd center of membership functionIs processed to generateAnd
(1) each row in the matrix T has at most one element with the value of 1, and the values of the other elements are 0;
(2) each column in the matrix T has only one element with the value of 1, and the values of the other elements are all 0;
(3) if all the elements of the ith row in the matrix T are 0, indicating that the ith reader in the original RFID reader fault self-adaptive positioning system has a fault;
since the transform matrix T is a sparse logic matrix, an array of size 1 × M' is created to store the row number i, i ═ 1,2ij=1。
Further, step S4 specifically includes the following steps:
s4.1, acquiring the signal intensity and the position data set of the RFID reference tag in real time;
s4.2, initializing the online sequential fuzzy width learning system in an offline stage;
and S4.3, updating the initial online sequence fuzzy width learning system obtained in the offline stage in the online stage and completing positioning.
Further, in step S4.1, the signal strength of the RFID reference tag is obtained in real time by controlling the read-write state of the RFID reader, and the corresponding position information of the reference tag in the system is obtained by the RFID reader fault adaptive positioning system.
Further, step S4.2 comprises the steps of:
s4.2.1, preprocessing the data obtained in the step S4.1;
s4.2.2 at [0,1 ]]In-range, randomly generating fuzzy subsystems according to uniform distributionCoefficient of function
S4.2.3, calculating the intermediate output of all input samples in the fuzzy subsystem in the ith fuzzy subsystemAnd an output
S4.2.4 according to the formulaComputing the intermediate outputs of all n fuzzy subsystemsAs input to the enhancement node;
s4.2.5 according to the formulaAnd formulaComputing output of jth group of enhanced nodesAnd total output of the enhancement layerWhereinHeelAs intermediate output of the fuzzy subsystemThe weight and bias of the connection with the enhanced node group are set at 0,1]Randomly generating within the range, wherein m is the group number of the enhanced nodes;
s4.2.6 according to the formulaCalculating the total output of all n fuzzy subsystemsWhereinΩ=(Ω1,...,Ωn),Δ=((δ1)T,...,(δn)T)T,Andin order to input the data, the data is,is the weighted output of the fuzzy rule and,for the parameters introduced, the initial ones can be usedIs converted intoWhen the solution is carried out through the pseudo-inverse in the subsequent steps, the weight number of the solution is determined byBecome into
S4.2.7 according to the formulaComputing the matrix A0And a parameter matrix W0Wherein W iseFor the connection weights of the enhancement layer to the output layer, is A0The calculation formula of (A) is
Further, step S4.2.1 includes the steps of:
s4.2.1.1, the firstRepeatedly reading the signal intensity of the same label by each reader for N times, and recording the signal intensity read at the kth time as
S4.2.1.2, calculating the variance delta of RSSI value2:
S4.2.1.3 for the kth signal strengthIf it is notThen removing the RSSI set with the size of N', and calculating the average value of the setAs the average signal intensity:
Step S4.2.3 includes the steps of:
s4.2.3.1, input data X using k-means clustering0Clustering is carried out to obtain KiA cluster center;
s4.2.3.2, use K obtained in step S4.2.3.1iCenter of individual clustering center to Gaussian membership functionInitialising, width of membership function
S4.2.3.3 according to the formulaAnd formulaComputing an intermediate output of an s-th input sample in a fuzzy subsystemAnd an output
S4.2.3.4 according to the formulaAnd formulaComputing intermediate outputs of all input samples in a fuzzy subsystemAnd an output
Further, in step S4.3, the online stage includes an online sequential learning stage and an online working stage, the online sequential learning stage is used for completing updating of the initial online sequential fuzzy width learning system, the online working stage is used for completing positioning, and the online sequential learning stage and the online working stage can run in parallel;
in the online sequential learning stage, the number of readers is changed from M to M' and 0<M' is less than or equal to M, and newly added data areThe method specifically comprises the following steps:
s4.3.1.1, preprocessing the data using step S4.2.1;
s4.3.1.2, generating a transformation matrix T according to the rule of the transformation matrix T;
s4.3.1.3 according to the formulaAnd formulaCalculating transformed coefficientsAnd center of membership function
S4.3.1.4 according to the formulaCalculating the operation output matrix A of the newly added dataaWherein Andin order to add new input data to the data,intermediate output of all newly added input samples in the fuzzy subsystem;
Further, the online working phase specifically includes the following steps:
s4.3.2.1, sending the RSSI information of the target label to an online sequence fuzzy width learning system after being updated in an online sequence learning stage;
s4.3.2.2, the position of the target label is estimated by using the RSSI information as the input of the online sequential fuzzy width learning system after the updating of the online sequential learning stage.
Compared with the prior art, the invention has the following advantages and technical effects:
the core of the method is that the signal intensity and the position of a reference label are firstly obtained in an off-line stage, a fuzzy subsystem and an enhancement layer of a fuzzy width system are initialized after data preprocessing is carried out, and finally an initial positioning model is obtained through training; in the online sequence learning stage, firstly, the new data of the RFID readers with the changed number are preprocessed, then, the coefficients and the membership centers are converted according to the conversion matrix, and finally, the initial positioning model obtained by the off-line stage training is updated according to the updating function; in the on-line working stage, firstly, the data of the positioning request sent by the user is preprocessed, then the preprocessed data is input into the updated model, and the positioning result is calculated and sent to the client. By the aid of the RFID reader fault self-adaptive positioning method, data streams which continuously arrive in the environment can be processed, and subsequent data streams can be processed when partial readers are in fault, so that the aim of long-time stable positioning is achieved.
Drawings
FIG. 1 is a schematic diagram of a deployment of an RFID reader fault adaptive location system in an embodiment of the present invention;
FIG. 2 is a schematic diagram of a model structure of a fuzzy width learning system according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an expanded structure of a fuzzy subsystem of the fuzzy width learning system according to the embodiment of the present invention;
FIG. 4 is a schematic diagram of network update of an online sequential fuzzy width learning system according to an embodiment of the present invention;
FIG. 5 illustrates an exemplary RFID reader failure in an embodiment of the invention;
FIG. 6 is a schematic diagram of an RFID reader fault adaptation strategy in an embodiment of the present invention;
FIG. 7 is a flowchart of an RFID reader fault adaptive location method according to an embodiment of the present invention.
Detailed Description
In order to make the technical solutions and advantages of the present invention more apparent, the following detailed description of the embodiments of the present invention is provided with reference to the accompanying drawings and examples, but the embodiments and protection of the present invention are not limited thereto.
Example (b):
an RFID reader fault self-adaptive positioning method comprises the following steps:
s1, constructing a fault self-adaptive positioning system of the RFID reader;
as shown in fig. 1, the RFID reader fault adaptive positioning system includes M RFID readers, Q RFID reference tags, and a target tag to be positioned; the reference tags with known coordinate positions are deployed in a plane where the target tags to be positioned are located in an equilateral triangle mode, and the readers are deployed on four edges of the plane.
S2, providing an online sequence fuzzy width learning system according to the RFID reader fault self-adaptive positioning system;
the model structure of the fuzzy width learning system is shown in fig. 2, wherein the fuzzy subsystem development structure is shown in fig. 3, and an online sequential fuzzy width learning system is proposed by improving the fuzzy width learning system, as shown in fig. 4The learning system comprises an input layer, a fuzzy subsystem, an enhancement layer and an output layer, wherein the fuzzy subsystem is a TS fuzzy model, and in the ith fuzzy subsystem, the weightWherein xsMFor input data, s1, 2, N is the number of input data, M is the input data dimension, KiIn order to be able to blur the number of rules,for randomly generated coefficients, the membership function is a Gaussian functionWhereinRespectively, the center and width of the membership function, and the weighted output of the fuzzy rule isWhereinThe enhancement layer is a nonlinear transformation layer;
the number of the fuzzy subsystems is n, the number of the enhanced node groups is m, wherein the fuzzy rule number of the ith fuzzy subsystem is KiThe number of the j-th group of enhanced nodes is LjOf 1 atThe signal strength of the RFID reader reading the reference label at the s time is Generation of N0The bar input data isThe output data isWhereinIs the Nth0An array of signal strengths of the strip input data,is the Nth0Coordinates of the input data are obtained, and C is a characteristic dimension of the output data;
will N0The bar input data is used as the input of a fuzzy subsystem in the online sequential fuzzy width learning system, the fuzzy subsystem intermediate output and the fuzzy subsystem output are obtained through calculation, the fuzzy subsystem intermediate output is used as the input of an enhancement node, the enhancement layer output is further obtained through calculation, the fuzzy subsystem output and the enhancement layer output are connected with the output layer, and the weight to be calculated is obtained through pseudo-inverse.
S3, providing a fault self-adaptive strategy of the RFID reader based on the online sequence fuzzy width learning system;
as shown in FIG. 5, when some of the RFID readers fail, the number of the RFID readers becomes M's, 0<M' is less than or equal to M, then subsequently generated NaThe bar input data isRSSIs=(RSSIs1,RSSIs2,...,RSSIsM′),s=1,2,...,NaThe output data isWhereinIs the NthaSignal Strength array, RSSI, of the strip input datasM′For the signal strength of the mth reader reading the reference tag at the s-th time,is the NthaCoordinates of the bar input data;
after partial RFID readers are in fault, the characteristic dimension of input data is changed from M to M', and a conversion matrix T needs to be introduced into an initialization fuzzy subsystemCoefficients generated randomly by a functionAnd center of membership functionIs processed to generateAnd
(1) each row in the matrix T has at most one element with the value of 1, and the values of the other elements are 0;
(2) each column in the matrix T has only one element with the value of 1, and the values of the other elements are all 0;
(3) if all the elements of the ith row in the matrix T are 0, indicating that the ith reader in the original RFID reader fault self-adaptive positioning system has a fault;
since the transform matrix T is a sparse logic matrix, an array of size 1 × M' is created to store the row number i, i ═ 1,2ij=1。
S4, completing RFID reader fault self-adaptive positioning based on the RFID reader fault self-adaptive strategy; as shown in fig. 7, the method specifically includes the following steps:
s4.1, acquiring the signal intensity and the position data set of the RFID reference tag in real time; the method comprises the following steps:
the signal strength of the RFID reference label is obtained in real time by controlling the read-write state of the RFID reader, and the corresponding position information of the reference label in the system is obtained by the RFID reader fault self-adaptive positioning system.
S4.2, initializing the online sequential fuzzy width learning system in an offline stage; the method comprises the following steps:
s4.2.1, preprocessing the data obtained in the step S4.1; the method comprises the following steps:
s4.2.1.1, the firstRepeatedly reading the signal intensity of the same label by each reader for N times, and recording the signal intensity read at the kth time as
S4.2.1.2, calculating the variance delta of RSSI value2:
S4.2.1.3 for the kth signal strengthIf it is notThen removing is carried out to obtain the size N′Calculating the mean of the setAs the average signal intensity:
S4.2.2 at [0,1 ]]In-range, randomly generating fuzzy subsystems according to uniform distributionCoefficient of function
S4.2.3, calculating the intermediate output of all input samples in the fuzzy subsystem in the ith fuzzy subsystemAnd an output The method comprises the following steps:
s4.2.3.1, input data X using k-means clustering0Clustering is carried out to obtain KiA cluster center;
s4.2.3.2, use K obtained in step S4.2.3.1iIndividual cluster centerTo center of Gaussian membership functionInitialising, width of membership function
S4.2.3.3 according to the formulaAnd formulaComputing an intermediate output of an s-th input sample in a fuzzy subsystemAnd an output
S4.2.3.4 according to the formulaAnd formulaComputing intermediate outputs of all input samples in a fuzzy subsystemAnd an output
S4.2.4 according to the formulaComputing the intermediate outputs of all n fuzzy subsystemsTo act as an enhanced nodeInputting;
s4.2.5 according to the formulaAnd formulaComputing output of jth group of enhanced nodesAnd total output of the enhancement layerWhereinHeelAs intermediate output of the fuzzy subsystemThe weight and bias of the connection with the enhanced node group are set at 0,1]Randomly generating within the range, wherein m is the group number of the enhanced nodes;
s4.2.6 according to the formulaCalculating the total output of all n fuzzy subsystemsWhereinΩ=(Ω1,...,Ωn),Δ=((δ1)T,...,(δn)T)T,Andin order to input the data, the data is,is the weighted output of the fuzzy rule and,for the parameters introduced, the initial ones can be usedIs converted intoWhen the solution is carried out through the pseudo-inverse in the subsequent steps, the weight number of the solution is determined byBecome into
S4.2.7 according to the formulaComputing the matrix A0And a parameter matrix W0Wherein W iseFor the connection weights of the enhancement layer to the output layer,is A0The calculation formula of (A) is
S4.3, updating and positioning the initial online sequence fuzzy width learning system obtained in the offline stage in the online stage;
the online stage comprises an online sequential learning stage and an online working stage, the online sequential learning stage is used for completing updating of the initial online sequential fuzzy width learning system, the online working stage is used for completing positioning, and the online sequential learning stage and the online working stage can run in parallel;
in the online sequential learning stage, the number of readers is changed from M to M′,0<M′Less than or equal to M, newly added data isThe method specifically comprises the following steps:
s4.3.1.1, preprocessing the data using step S4.2.1;
s4.3.1.2, generating a transformation matrix T according to the rule of the transformation matrix T;
s4.3.1.3 according to the formulaAnd formulaCalculating transformed coefficientsAnd center of membership function
S4.3.1.4 according to the formulaCalculating the operation output matrix A of the newly added dataaWherein Andin order to add new input data to the data,intermediate output of all newly added input samples in the fuzzy subsystem;
The online working phase specifically comprises the following steps:
s4.3.2.1, sending the RSSI information of the target label to an online sequence fuzzy width learning system after being updated in an online sequence learning stage;
s4.3.2.2, the position of the target label is estimated by using the RSSI information as the input of the online sequential fuzzy width learning system after the updating of the online sequential learning stage.
Claims (10)
1. A self-adaptive fault positioning method for an RFID reader is characterized by comprising the following steps:
s1, constructing a fault self-adaptive positioning system of the RFID reader;
s2, providing an online sequence fuzzy width learning system according to the RFID reader fault self-adaptive positioning system;
s3, providing a fault self-adaptive strategy of the RFID reader based on the online sequence fuzzy width learning system;
s4, completing RFID reader fault adaptive positioning based on the RFID reader fault adaptive strategy.
2. The RFID reader fault self-adaptive positioning method according to claim 1, wherein in step S1, the RFID reader fault self-adaptive positioning system comprises M RFID readers, Q RFID reference tags and a target tag to be positioned; the reference tags with known coordinate positions are deployed in a plane where the target tags to be positioned are located in an equilateral triangle mode, and the readers are deployed on four edges of the plane.
3. The RFID reader fault self-adaptive positioning method of claim 1, wherein in step S2, the online sequential fuzzy width learning system comprises an input layer, a fuzzy subsystem, an enhancement layer and an output layer, wherein the fuzzy subsystem is a TS fuzzy model, and in the ith fuzzy subsystem, the weight is used as a weight Wherein xsMFor input data, s1, 2, N is the number of input data, M is the input data dimension, KiIn order to be able to blur the number of rules,for randomly generated coefficients, the membership function is a Gaussian function WhereinRespectively, the center and width of the membership function, and the weighted output of the fuzzy rule isWhereinThe enhancement layer is a nonlinear transformation layer;
the number of the fuzzy subsystems is n, the number of the enhanced node groups is m, wherein the fuzzy rule number of the ith fuzzy subsystem is KiThe number of the j-th group of enhanced nodes is LjOf 1 atThe signal strength of the RFID reader reading the reference label at the s time is Generation of N0The bar input data isRSSIs=(RSSIs1,RSSIs2,...,RSSIsM),s=1,2,...,N0The output data isWhereinIs the Nth0An array of signal strengths of the strip input data,is the Nth0Coordinates of the input data are obtained, and C is a characteristic dimension of the output data;
will N0The bar input data is used as the input of a fuzzy subsystem in the online sequential fuzzy width learning system, and the intermediate output of the fuzzy subsystem is obtained through calculationAnd fuzzy subsystem output, wherein the middle output of the fuzzy subsystem is used as the input of the enhancement node, the enhancement layer output is obtained through further calculation, the fuzzy subsystem output and the enhancement layer output are connected with the output layer, and the weight to be calculated is obtained through pseudo-inverse.
4. The method according to claim 1, wherein in step S3, the adaptive strategy for RFID reader fault is as follows:
when some RFID readers are in fault, the number of the RFID readers is changed into M' and 0<M' is less than or equal to M, then subsequently generated NaThe bar input data isRSSIs=(RSSIs1,RSSIs2,...,RSSIsM′),s=1,2,...,NaThe output data isWhereinIs the NthaSignal Strength array, RSSI, of the strip input datasM′For the signal strength of the mth reader reading the reference tag at the s-th time,is the NthaCoordinates of the bar input data;
after partial RFID readers are in fault, the characteristic dimension of input data is changed from M to M', and a conversion matrix T needs to be introduced into an initialization fuzzy subsystemCoefficients generated randomly by a functionAnd center of membership functionIs processed to generateAnd
(1) each row in the matrix T has at most one element with the value of 1, and the values of the other elements are 0;
(2) each column in the matrix T has only one element with the value of 1, and the values of the other elements are all 0;
(3) if all the elements of the ith row in the matrix T are 0, indicating that the ith reader in the original RFID reader fault self-adaptive positioning system has a fault;
since the transform matrix T is a sparse logic matrix, an array of size 1 × M' is created to store the row number i, i ═ 1,2ij=1。
5. The self-adaptive fault location method for the RFID reader according to claim 1, wherein the step S4 specifically comprises the following steps:
s4.1, acquiring the signal intensity and the position data set of the RFID reference tag in real time;
s4.2, initializing the online sequential fuzzy width learning system in an offline stage;
and S4.3, updating the initial online sequence fuzzy width learning system obtained in the offline stage in the online stage and completing positioning.
6. The self-adaptive fault location method for the RFID reader as claimed in claim 5, wherein in step S4.1, the signal strength of the RFID reference tag is obtained in real time by controlling the read-write state of the RFID reader, and the corresponding position information of the reference tag in the system is obtained by the self-adaptive fault location system for the RFID reader.
7. The RFID reader fault self-adaptive positioning method according to claim 5, characterized in that the step S4.2 comprises the following steps:
s4.2.1, preprocessing the data obtained in the step S4.1;
s4.2.2 at [0,1 ]]In-range, randomly generating fuzzy subsystems according to uniform distributionCoefficient of function
S4.2.3, calculating the intermediate output of all input samples in the fuzzy subsystem in the ith fuzzy subsystemAnd an outputi=1,2,...,n;
S4.2.4 according to the formulaComputing the intermediate outputs of all n fuzzy subsystemsAs input to the enhancement node;
s4.2.5 according to the formulaAnd formulaComputing output of jth group of enhanced nodesAnd total output of the enhancement layerWhereinHeelAs intermediate output of the fuzzy subsystemThe weight and bias of the connection with the enhanced node group are set at 0,1]Randomly generating within the range, wherein m is the group number of the enhanced nodes;
s4.2.6 according to the formulaCalculating the total output of all n fuzzy subsystemsWhereinΩ=(Ω1,...,Ωn),Δ=((δ1)T,...,(δn)T)T, In order to input the data, the data is,is the weighted output of the fuzzy rule and,for the parameters introduced, the initial ones can be usedIs converted intoWhen the solution is carried out through the pseudo-inverse in the subsequent steps, the weight number of the solution is determined byBecome into
8. The method of claim 7, wherein the step S4.2.1 comprises the following steps:
s4.2.1.1, the firstRepeatedly reading the signal intensity of the same label by each reader for N times, and recording the signal intensity read at the kth time as
S4.2.1.2, calculating the variance delta of RSSI value2:
S4.2.1.3 for the kth signal strengthIf it is notThen removing the RSSI set with the size of N', and calculating the average value of the setAs the average signal intensity:
Step S4.2.3 includes the steps of:
s4.2.3.1, input data X using k-means clustering0Clustering is carried out to obtain KiA cluster center;
s4.2.3.2, use K obtained in step S4.2.3.1iCenter of individual clustering center to Gaussian membership functionInitialising, width of membership function
S4.2.3.3 according to the formulaAnd formulaComputing an intermediate output of an s-th input sample in a fuzzy subsystemAnd an outputs=1,2,...,N0;
9. The RFID reader fault self-adaptive positioning method according to claim 5, characterized in that in step S4.3, the online stage comprises an online sequential learning stage and an online working stage, the online sequential learning stage is used for completing updating of an initial online sequential fuzzy width learning system, the online working stage is used for completing positioning, and the online sequential learning stage and the online working stage can run in parallel;
in the online sequential learning stage, the number of readers is changed from M to M' and 0<M' is less than or equal to M, and newly added data areThe method specifically comprises the following steps:
s4.3.1.1, preprocessing the data using step S4.2.1;
s4.3.1.2, generating a transformation matrix T according to the rule of the transformation matrix T;
s4.3.1.3 according to the formulaAnd formulaCalculating transformed coefficientsAnd center of membership function
S4.3.1.4 according to the formulaCalculating the operation output matrix A of the newly added dataaWherein And in order to add new input data to the data,intermediate output of all newly added input samples in the fuzzy subsystem;
10. The method according to claim 9, wherein the online working phase specifically comprises the following steps:
s4.3.2.1, sending the RSSI information of the target label to an online sequence fuzzy width learning system after being updated in an online sequence learning stage;
s4.3.2.2, the position of the target label is estimated by using the RSSI information as the input of the online sequential fuzzy width learning system after the updating of the online sequential learning stage.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010028807.1A CN111199162B (en) | 2020-01-11 | 2020-01-11 | RFID reader fault self-adaptive positioning method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010028807.1A CN111199162B (en) | 2020-01-11 | 2020-01-11 | RFID reader fault self-adaptive positioning method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111199162A true CN111199162A (en) | 2020-05-26 |
CN111199162B CN111199162B (en) | 2021-10-26 |
Family
ID=70747267
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010028807.1A Active CN111199162B (en) | 2020-01-11 | 2020-01-11 | RFID reader fault self-adaptive positioning method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111199162B (en) |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101540014A (en) * | 2008-03-17 | 2009-09-23 | 大叶大学 | Information system for applying radio frequency identification to facility and equipment maintenance and management and method thereof |
CN101587182A (en) * | 2009-06-25 | 2009-11-25 | 华南理工大学 | Locating method for RFID indoor locating system |
CN101639527A (en) * | 2009-09-03 | 2010-02-03 | 哈尔滨工业大学 | K nearest fuzzy clustering WLAN indoor locating method based on REE-P |
CN106555788A (en) * | 2016-11-11 | 2017-04-05 | 河北工业大学 | Application of the deep learning based on Fuzzy Processing in hydraulic equipment fault diagnosis |
CN106597425A (en) * | 2016-11-18 | 2017-04-26 | 中国空间技术研究院 | Radar object positioning method based on machine learning |
WO2018111601A1 (en) * | 2016-12-16 | 2018-06-21 | Square, Inc. | Tamper detection system |
US20190065939A1 (en) * | 2017-08-30 | 2019-02-28 | International Business Machines Corporation | Bayesian network based hybrid machine learning |
CN109598320A (en) * | 2019-01-16 | 2019-04-09 | 广西大学 | A kind of RFID indoor orientation method based on locust algorithm and extreme learning machine |
CN109816068A (en) * | 2019-01-28 | 2019-05-28 | 太原理工大学 | A kind of detection method of the mobile tag based on radio-frequency recognition system |
CN110363165A (en) * | 2019-07-18 | 2019-10-22 | 深圳大学 | Multi-object tracking method, device and storage medium based on TSK fuzzy system |
-
2020
- 2020-01-11 CN CN202010028807.1A patent/CN111199162B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101540014A (en) * | 2008-03-17 | 2009-09-23 | 大叶大学 | Information system for applying radio frequency identification to facility and equipment maintenance and management and method thereof |
CN101587182A (en) * | 2009-06-25 | 2009-11-25 | 华南理工大学 | Locating method for RFID indoor locating system |
CN101639527A (en) * | 2009-09-03 | 2010-02-03 | 哈尔滨工业大学 | K nearest fuzzy clustering WLAN indoor locating method based on REE-P |
CN106555788A (en) * | 2016-11-11 | 2017-04-05 | 河北工业大学 | Application of the deep learning based on Fuzzy Processing in hydraulic equipment fault diagnosis |
CN106597425A (en) * | 2016-11-18 | 2017-04-26 | 中国空间技术研究院 | Radar object positioning method based on machine learning |
WO2018111601A1 (en) * | 2016-12-16 | 2018-06-21 | Square, Inc. | Tamper detection system |
US20190065939A1 (en) * | 2017-08-30 | 2019-02-28 | International Business Machines Corporation | Bayesian network based hybrid machine learning |
CN109598320A (en) * | 2019-01-16 | 2019-04-09 | 广西大学 | A kind of RFID indoor orientation method based on locust algorithm and extreme learning machine |
CN109816068A (en) * | 2019-01-28 | 2019-05-28 | 太原理工大学 | A kind of detection method of the mobile tag based on radio-frequency recognition system |
CN110363165A (en) * | 2019-07-18 | 2019-10-22 | 深圳大学 | Multi-object tracking method, device and storage medium based on TSK fuzzy system |
Also Published As
Publication number | Publication date |
---|---|
CN111199162B (en) | 2021-10-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Schäfer et al. | Recurrent neural networks are universal approximators | |
Zou et al. | An RFID indoor positioning system by using weighted path loss and extreme learning machine | |
CN109212476B (en) | RFID indoor positioning algorithm based on DDPG | |
Li et al. | TransLoc: A heterogeneous knowledge transfer framework for fingerprint-based indoor localization | |
Zhu et al. | BLS-location: A wireless fingerprint localization algorithm based on broad learning | |
Zhang et al. | Systematic comparison of graph embedding methods in practical tasks | |
Zhong et al. | RF-OSFBLS: An RFID reader-fault-adaptive localization system based on online sequential fuzzy broad learning system | |
CN111050282A (en) | Multi-time fuzzy inference weighted KNN positioning method | |
CN111050294A (en) | Indoor positioning system and method based on deep neural network | |
Xu et al. | Neural network-based accuracy enhancement method for WLAN indoor positioning | |
Ertuğrul et al. | Fuzzy TOPSIS method for academic member selection in engineering faculty | |
CN111199162B (en) | RFID reader fault self-adaptive positioning method | |
Trinh et al. | Bearing-based formation control and network localization via global orientation estimation | |
Zhu et al. | Prediction of battlefield target trajectory based on LSTM | |
Yan et al. | An ELM-based semi-supervised indoor localization technique with clustering analysis and feature extraction | |
CN109766969B (en) | RFID indoor positioning algorithm based on asynchronous dominant motion evaluation | |
Fan et al. | WiFi based indoor localization with multiple kernel learning | |
CN110691319A (en) | Method for realizing high-precision indoor positioning of heterogeneous equipment in self-adaption mode in use field | |
CN113052219B (en) | Abnormal track detection method and device and electronic equipment | |
Yadav et al. | Research on indoor positioning technology of RFID nodes in the internet of things (IOT) | |
Yan et al. | RFID positioning algorithm based on BA optimization | |
Sobhiyeh et al. | Online detection and parameter estimation with correlated data in wireless sensor networks | |
Singh et al. | Hybrid optimization algorithm for community and fraud detection in complex networks for high immunity towards link and node failures | |
CN114745658B (en) | Distance estimation method and medium based on decision fusion | |
CN110888109B (en) | RFID label positioning method based on generalized multidimensional scale |
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 |