CN111225367B - RFID network planning method based on hybrid particle swarm optimization - Google Patents

RFID network planning method based on hybrid particle swarm optimization Download PDF

Info

Publication number
CN111225367B
CN111225367B CN202010018027.9A CN202010018027A CN111225367B CN 111225367 B CN111225367 B CN 111225367B CN 202010018027 A CN202010018027 A CN 202010018027A CN 111225367 B CN111225367 B CN 111225367B
Authority
CN
China
Prior art keywords
reader
particles
population
particle
readers
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010018027.9A
Other languages
Chinese (zh)
Other versions
CN111225367A (en
Inventor
刘静
曹雅婷
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.)
Xidian University
Original Assignee
Xidian University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xidian University filed Critical Xidian University
Priority to CN202010018027.9A priority Critical patent/CN111225367B/en
Publication of CN111225367A publication Critical patent/CN111225367A/en
Application granted granted Critical
Publication of CN111225367B publication Critical patent/CN111225367B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/80—Services using short range communication, e.g. near-field communication [NFC], radio-frequency identification [RFID] or low energy communication
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/18—Network planning tools
    • H04W16/20—Network planning tools for indoor coverage or short range network deployment
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition

Abstract

The invention discloses a RFID network planning method based on a hybrid particle swarm algorithm, which mainly solves the problems that the number of readers deployed in the RFID network planning in the prior art is too many and the mutual interference is large, and has the scheme that: simulating particle swarm movement to construct a wireless radio frequency network RFID system model, determining the number of readers of the model, randomly obtaining the initial position and the power radius of each reader as a particle, and repeatedly obtaining a population; setting particle swarm algorithm parameters and evaluating various performances of particles in a population; recording the individual optimal particles and the population optimal particles in a layered mode and updating the position of each particle; adjusting the position of the reader in each particle according to the covering state of the label; and obtaining the optimal individual which is the RFID network planning scheme after multiple iterations. The invention can quickly obtain the optimization result, reduce the number of readers in the network and avoid the interference between the readers, and can be used for the engineering RFID network optimization.

Description

RFID network planning method based on hybrid particle swarm optimization
Technical Field
The invention belongs to the technical field of networks, and particularly relates to a Chinese RFID network planning method which can be used for logistics, traffic, identity recognition, anti-counterfeiting, asset management, food and information statistics, reference application and safety control.
Background
The Radio Frequency Identification (RFID) technology is a non-contact automatic identification technology. The application of radio frequency identification is very wide, and at present, typical application fields include logistics, traffic, identity identification, anti-counterfeiting, asset management, food, information statistics, reference application and safety control. In the rfid system, an electronic tag exchanges data with a reader through radio waves. Because the detection range of the reader is limited and the number of the tags is large, in order to ensure the communication quality, setting the appropriate number of the readers and deploying the readers is a very important technical problem. If the number of readers is small, the system cannot detect all the tags; when the number of readers is too large and the coverage areas overlap, interference occurs therebetween. All of which affect the communication between the reader and the tag. Meanwhile, the total power and the load balance of the RFID system are also the performances that should be considered in the optimization process. More and more scholars design different algorithms to optimize the RFID network and determine the number and the positions of readers to be deployed in the network. In the practical application of the RFID system, the existing research results still have great defects. In order to further improve the communication quality, enhance the system stability and save the cost, a reasonably designed RFID network planning method is still necessary and feasible.
S.S. Shide et al, "Multi-Objective evolution Algorithm Based application for Solving RFID Reader plan Using Weight-vector with application-Based Learning Method" ("International Journal of Record Technology and Engineering (IJRTE), article No. 2277 3878(2019)), the core of the article is to search the solution space by Using a Multi-Objective Algorithm, specifically: and taking the maximized tag coverage rate, the minimized interference rate and the minimized cost as fitness functions, adding weights to the fitness functions, initializing the population by adopting an opposition-based learning method, and executing a multi-objective differential evolution algorithm. The individuals with the highest fitness value are saved in each generation until a termination condition is met. The disadvantages of this algorithm are: the number of readers needs to be estimated in advance before searching, and the reasonable number of readers has great influence on the RFID network planning result. In practical engineering applications, the number of readings to be set in a network is often difficult to determine. For the RFID system, the performance indexes have clear priorities, if a plurality of objective functions are considered simultaneously in a weighting mode, the most critical objective function is difficult to achieve the optimal solution, and the RFID system is difficult to be applied to actual engineering.
Zhang et al, "An effective and fast mechanics-based algorithm for RFID network planning" ("Computer Networks", article number: 13-24(2017)), the core of the article is to simulate curling movement to search solution space, specifically, each reader is regarded as a curling capable of sliding in a working area, and a label is used for generating friction to prevent the reader from moving; and designing a mobile operator and a collision operator of the reader to perform searching operation, and reducing the number of the readers to be deployed in the network by combining with a redundant reader elimination strategy. Although the method can adaptively determine the number of the readers in the searching process, the interference rate, the total power and the load balance degree of the RFID system are not directly considered, the priority of each performance is also ignored, and in addition, the introduction of redundant reader elimination strategies makes four optimization results of the system not optimal, and the realization of actual engineering is difficult.
Disclosure of Invention
The invention aims to provide an RFID network planning method based on a hybrid particle swarm algorithm aiming at the defects of the prior art, so that the comprehensive performance of an RFID system can be optimized under the condition of introducing no redundant reader.
The technical scheme of the invention is as follows: the method comprises the following steps of adopting a K-means algorithm to adaptively determine the number of readers in a network, optimizing coverage, interference rate, total power and load balance degree in a layering mode by combining with the priority of the performance of an RFID system, introducing virtual force in the optimization process of a particle swarm algorithm, adaptively determining the number of readers in the network through label distribution in the network, and obtaining a reasonable reader position deployment result according to the priority of the performance of the RFID system, wherein the implementation steps comprise the following steps:
(1) simulating particle swarm motion to construct a wireless radio frequency network RFID system model, and respectively defining the distribution of a network working area, a tag and a reader;
(2) defining the initial value of the number of readers in the network as: n is 3;
(3) classifying the labels in the network by using a K mean value algorithm to obtain m readers and a label classification result TS of each readerjWhere j is ∈ [0, m ]]；
(4) Calculating the power radius u of each readerjAnd setting a threshold range T ═ epsilon of the power radius according to the requirements in practical engineeringmin,εmax]In which epsilonminIs the minimum value of the available reader power radius, epsilonmaxThe maximum value of the available reader power radius;
(5) judgment ujWhether threshold range T is satisfied:
if ujIn the threshold rangeWithin T, determining the number of readers in the network as m;
if ujIf the number m of the readers in the network is not within the threshold range T, the number m of the readers in the network is adjusted to be m +1, and the step (3) of reclassifying is returned;
(6) according to the number of readers in a network, randomly obtaining the initial position and the power radius of each reader through a K-means algorithm, taking the positions and the power radii of all the readers as a particle, and repeatedly executing N times to obtain N particles as a population;
(7) evaluating each performance of particles in the population by using four fitness functions, namely a coverage rate COV, an interference rate INT, a total power POW and a load balance degree LB in sequence;
(8) recording the optimal value of each particle, namely the individual optimal particle, and the optimal value of all particles in the population, namely the global optimal particle, in a layered mode according to the evaluation result of the step (7);
(9) setting the iteration times of the population as gen, and updating the speed and the position of the particles in an iterative manner by utilizing a particle swarm algorithm updating strategy according to the individual optimal particles and the global optimal particles obtained in the step (8);
(10) and adjusting the position of the reader in each particle according to the covering state of the label:
if the label is covered by a plurality of readers at the same time, repeatedly selecting any one of the two readers with the nearest distance for position adjustment, so that the coordinate position changes after mutual repulsion between the readers until the label is covered by only one reader;
if the label is not covered by any reader, the reader closest to the label is moved to the position of the label, and then the coordinate is changed until the label is covered by the reader;
(11) and (5) repeatedly executing the steps (7) to (10) until the maximum population algebra times gen are met, and obtaining the global optimal particles, wherein the optimal particles are the RFID network planning scheme, namely the deployment position and the power of the reader.
Compared with the prior art, the method has the following advantages:
firstly, the number of readers in the network is sequentially increased at the initial stage, and the labels in the network are classified and the power radius is calculated through a K mean value algorithm until the power radius meets the engineering requirement, so that the number of readers in the network can be determined in a self-adaptive manner; meanwhile, in the implementation process of the invention, the number of the readers is sequentially increased, rather than a large number of readers are deployed in the initialization process, so that the operation greatly reduces the calculation amount.
Secondly, after the number of readers in the RFID network is determined, the method adopts the K-means algorithm for initialization, so that compared with other random search methods, the method reduces the search space of the optimal solution, reduces the calculated amount and improves the speed of searching the optimal solution.
Thirdly, in the optimization process of the particle swarm optimization, the position of the reader in each particle is adjusted according to the coverage state of the label, so that the quality of the optimal solution is effectively improved, the convergence speed is accelerated, and the position of the reader can be optimized more simply and effectively.
Fourthly, because the invention utilizes a layered mode to evaluate the four performances of the RFID system in turn, rather than optimizing the four performances in a weighting mode, the performance of the RFID system can be optimized more reasonably, and the performance of the RFID system is more suitable for practical application.
Drawings
FIG. 1 is a flow chart of an implementation of the present invention;
FIG. 2 is a diagram of a model RFID system of the present invention;
FIG. 3 is a schematic diagram illustrating coordinate position changes after mutual repulsion between readers according to the present invention;
FIG. 4 is a schematic diagram illustrating a change in coordinate position after a reader approaches a tag position in the present invention;
FIG. 5 is a graphical representation of the results of a simulation of different data sets using the present invention;
FIG. 6 is a comparison graph of the coverage rate versus the number of readers during the performance optimization of the RFID system using the present invention and the existing curling algorithm, MOEAD algorithm, respectively;
FIG. 7 is a comparison graph of the interference rate with the number of readers during the performance optimization process of the RFID system using the present invention and the existing curling algorithm and MOEAD algorithm, respectively.
Detailed Description
Embodiments and effects of the present invention are described in detail below with reference to the accompanying drawings.
Referring to fig. 1, the specific implementation steps of this embodiment are as follows:
step 1, simulating particle swarm motion to construct a wireless radio frequency network RFID system model.
As shown in fig. 2, the model of the wireless radio frequency network RFID system includes a plurality of readers and a plurality of tags, wherein the readers cover the tags in the power radius range and establish communication, and the tags are randomly distributed in the working area; and uploading the communication result of the reader and the tag to the RFID system host.
And 2, respectively defining the network working area, the label and the reader distribution.
2.1) defining the working area of the RFID network to be optimized as a particle swarm motion space, wherein the space boundary is 50m multiplied by 50 m;
2.2) the ith tag t to be coverediRandomly distributed in the working space in a uniform or cluster form, and the coordinates are expressed as
Figure BDA0002359657460000041
Where i ∈ [0, n ]]N is the number of tags in the network;
2.3) reader r of jjIs expressed as
Figure BDA0002359657460000042
j∈[0,m]And m is the number of readers in the network.
Step 3, classifying the labels in the network by using a K mean algorithm to obtain m readers and each reader rjIs classified into the label result TSj。
3.1) defining the initial value of the number of readers in the network as: m is 3;
3.2) randomly initializing the position coordinates of m readers
Figure BDA0002359657460000043
3.3) with each reader rjAs cluster center, assigning the label to the closest cluster to obtain each reader rjIs classified into the label result TSj，j∈[0,m]；
3.4) calculating the mean value of the label coordinates in each cluster to update the cluster center:
Figure BDA0002359657460000051
Figure BDA0002359657460000052
wherein the content of the first and second substances,
Figure BDA0002359657460000053
denotes the updated coordinates of the jth reader, (x)ij,yij) Indicates the ith distribution in the reader rjPosition coordinates of the label within the coverage, NtjIs shown in the reader rjThe number of tags within the coverage of (c);
3.5) updating the cluster center coordinates
Figure BDA0002359657460000054
From the original cluster center coordinates
Figure BDA0002359657460000055
And (3) comparison:
if the updated cluster center coordinates
Figure BDA0002359657460000056
From the original cluster center coordinates
Figure BDA0002359657460000057
If the two are equal, the classification is finished, and the step 4 is executed;
if the updated cluster center coordinates
Figure BDA0002359657460000058
From the original cluster center coordinates
Figure BDA0002359657460000059
Unequal, return 3.3).
Step 4, setting a threshold range T of the power radius, and calculating the power radius u of each readerj。
4.1) setting the threshold range of the power radius according to the requirements in practical engineering as follows: t ═ epsilonmin,εmax]In which epsilonminIs the minimum value of the available reader power radius, epsilonmaxThe maximum value of the available reader power radius;
4.2) calculating the power radius u of each readerj：
uj＝max(d(tij,rj)) tij∈TSj
Figure BDA00023596574600000510
Wherein r isjJ-th reader, u, representing a particlejIndicates the power radius of the jth reader,
Figure BDA00023596574600000511
indicating the position coordinates of the jth reader; t is tijIs shown in the reader rjThe ith tag in the coverage of (1),
Figure BDA00023596574600000512
indicating the location coordinates of the tag; d (t)ij,rj) Presentation reader rjAnd a label tijEuclidean distance of.
And 5, determining the number of readers in the network.
Radius of power u of readerjComparison with a threshold range T:
if ujWithin the threshold value range T, determining the number of readers in the network as m;
if ujIf the number m of the readers in the network is not within the threshold value range T, the number m of the readers in the network is adjusted to be m +1, and the step 3.2) is returned to reclassify.
And 6, according to the number m of the readers in the network, randomly obtaining the initial position and the power radius of each reader through a K-means algorithm, taking the positions and the power radii of all the readers as one particle, and repeatedly executing N times to obtain N particles as one population.
6.1) random initialization of the position coordinates of the m readers
Figure BDA0002359657460000061
6.2) with each reader rjAs cluster center, assigning the label to the closest cluster to obtain each reader rjIs classified into the label result TSj，j∈[0,m]；
6.3) calculating the mean value of the label coordinates in each cluster to update the cluster center:
Figure BDA0002359657460000062
Figure BDA0002359657460000063
wherein the content of the first and second substances,
Figure BDA0002359657460000064
denotes the updated coordinates of the jth reader, (x)ij,yij) Indicates the ith distribution in the reader rjPosition coordinates of the label within the coverage, NtjIs shown in the reader rjThe number of tags within the coverage of (c);
6.4) cluster center coordinates after updating
Figure BDA0002359657460000065
From the original cluster center coordinates
Figure BDA0002359657460000066
And (3) comparison:
if the updated cluster center coordinates
Figure BDA0002359657460000067
From the original cluster center coordinates
Figure BDA0002359657460000068
If the two readers are equal, finishing classification, obtaining the initial position of each reader and executing the step 3.5);
if the updated cluster center coordinates
Figure BDA0002359657460000069
From the original cluster center coordinates
Figure BDA00023596574600000610
Not equal, return 6.2).
6.5) calculating the Power radius u of each readerj：
uj＝max(d(tij,rj)) tij∈TSj
Figure BDA00023596574600000611
Wherein r isjJ-th reader, u, representing a particlejIndicates the power radius of the jth reader,
Figure BDA00023596574600000612
indicating the position coordinates of the jth reader; t is tijIs shown in the reader rjThe ith tag in the coverage of (1),
Figure BDA00023596574600000613
indicating the location coordinates of the tag; d (t)ij,rj) Presentation reader rjAnd a label tijEuclidean distance of.
Step 7, setting the iteration times of the population as gen, and the initial value of the population algebra: d is 0.
And 8, evaluating each performance of the particles in the population by using four fitness functions respectively.
8.1) calculating the particle coverage COV in the population by using an evaluation coverage fitness function:
Figure BDA0002359657460000071
wherein the content of the first and second substances,
Figure BDA0002359657460000072
ujdenotes the power radius of the jth reader, p (t)i,rj) Indicates the label tiIs read by the reader rjOver-coverage state of, NtIs the number of tags in the network, l (t)i,rj) Is a reader rjAnd a label tiEuclidean distance of.
8.2) calculating the inter-reader interference rate INT of the particles in the population by using the evaluation interference rate fitness function:
Figure BDA0002359657460000073
wherein
Figure BDA0002359657460000074
num(ti) To cover the label tiThe number of readers, rat (t)i) Is a label tiThe degree of interference of.
8.3) calculating the total power POW of the particles in the population by using an evaluation RFID system total power fitness function:
Figure BDA0002359657460000075
wherein RS represents a set of readers in the network;
8.4) calculating the load balance LB of the particles in the population by using the evaluation RFID system load balance fitness function:
Figure BDA0002359657460000076
wherein the content of the first and second substances,
Figure BDA0002359657460000077
is a reader rjThe number of labels covered.
And 9, recording the individual optimal particles and the global optimal particles in a layered mode according to the evaluation result of the step 8.
9.1) defining the individual optimal particles as the optimal value of each particle, and the global optimal particles as the optimal values of all the particles in the population, and judging whether the population is the first generation:
if the population is the first generation population, the individual optimal particle is the individual optimal particle;
if not, executing 9.2);
9.2) respectively comparing the coverage rate of the particles in the current population with the individual optimal particles:
if the two are equal, 9.3) is executed;
if the coverage rate of the particles in the current population is greater than that of the individual optimal particles, executing 9.6);
if the coverage rate of the particles in the current population is smaller than that of the individual optimal particles, ending;
9.3) respectively comparing the interference rates of the particles in the current population and the individual optimal particles:
if the interference rate of the particles in the current population is equal to the interference rate of the individual optimal particles, executing 9.4);
if the interference rate of the particles in the current population is less than the interference rate of the individual optimal particles, executing 9.6);
if the interference rate of the particles in the current population is greater than the interference rate of the individual optimal particles, ending;
9.4) respectively comparing the total system power of the particles in the current population with the individual optimal particles:
if the total system power of the particles in the current population is equal to the total system power of the individual optimal particles, executing 9.5);
if the total system power of the particles in the current population is less than that of the individual optimal particles, executing 9.6);
if the total system power of the particles in the current population is greater than that of the individual optimal particles, ending;
9.5) respectively comparing the load balance degree of the particles in the current population with the individual optimal particles:
if the load balance degree of the particles in the current population is smaller than the total system power of the individual optimal particles, executing 9.6);
if the load balance degree of the particles in the current population is more than or equal to the total system power of the individual optimal particles, ending;
9.6) updating the current particle to be the individual optimal particle, and performing hierarchical comparison on the four fitness values of the individual optimal particle and the global optimal particle respectively to update the individual optimal particle to be the global optimal particle.
And step 10, updating the positions of the particles in the next generation of population.
According to the individual optimal particles and the global optimal particles obtained in the step 9, updating a strategy by using a particle swarm algorithm, and calculating the positions of the particles in the next generation of population by the following iterative formula:
Figure BDA0002359657460000081
wherein the content of the first and second substances,
Figure BDA0002359657460000082
indicates the position of the ith particle in the population of the (d + 1) th generation,
Figure BDA0002359657460000083
denotes the position, V, of the ith particle in the population of the d generationi d+1Represents the velocity of the ith particle in the population of the (d + 1) th generation.
Figure BDA0002359657460000085
Vi dRepresenting the velocity of the ith particle in the current generation-d population,
Figure BDA0002359657460000087
represents the optimal position of the ith particle in the population of the d generation, gdRepresents the optimal position of all particles in the population of the d generation, c1，c2Representing two different learning factors, c1＝1，c2＝2，r1 d
Figure BDA0002359657460000088
Is between [0,1]And a random number different from the above, wherein omega represents an inertia factor, and omega is 0.5.
And 11, adjusting the position of the reader in each particle according to the covering state of the label.
11.1) judging whether the label is covered by the reader:
if the tag is covered by the reader, 11.2) is executed;
if the tag is not covered by any reader, then execute 11.4);
11.2) judging whether the label is covered by a reader:
if the tag is covered by a reader, executing step 12;
if the label is covered by a plurality of readers, executing 11.3);
11.3) repeatedly selecting any one of the two readers with the nearest distance, and adjusting the position of the reader by the following formula to change the coordinate position after mutual repulsion between the readers until the label is covered by only one reader, wherein the change process is as shown in fig. 3:
Figure BDA0002359657460000091
Figure BDA0002359657460000092
wherein the content of the first and second substances,
Figure BDA0002359657460000093
are respectively the first reader r1And a second reader r2A is between [0,1 ]]A random number in between, and a random number,
Figure BDA0002359657460000094
is a second reader r2Coordinates obtained after rejection;
11.4) the reader closest to the tag is moved to the position thereof, and then the coordinate changes until the tag is covered by the reader, wherein the change process is shown in fig. 4, and the calculation formula is as follows:
Figure BDA0002359657460000095
Figure BDA0002359657460000096
wherein the content of the first and second substances,
Figure BDA0002359657460000097
are respectively the first reader r1And a second reader r2Beta is between [0,1 ]]A random number in between, and a random number,
Figure BDA0002359657460000098
is a reader r2And (4) coordinates obtained after the label position is closed.
And step 12, obtaining the global optimal particles.
12.1) updating population generation numbers: d is d + 1;
12.2) comparing the current population generation number d with the maximum population generation number gen:
if the current population algebra d is smaller than the maximum population algebra times gen, returning to the step 8 to execute repeatedly until the current population algebra d reaches the maximum population algebra times gen;
if the current population algebra d is larger than or equal to the maximum population algebra times gen, obtaining the global optimal particles, wherein the optimal particles are the RFID network planning scheme, namely the deployment position and the power of the reader.
The effects of the present invention can be further illustrated by the following simulations:
1. simulation experiment conditions are as follows:
the hardware platform of the simulation experiment of the invention is as follows: the processor is an Intel i54210U CPU, the main frequency is 1.70GHz, and the memory is 4.00 GB.
The software platform of the simulation experiment of the invention is as follows: windows 10 operating system and Visual Studio 2017.
2. Setting experimental conditions:
the number of RFID test samples tested in the experiment is 8, which are respectively as follows: c _30, C _50, C _100, C _500, R _30, R _50, R _100, and R _ 500. Wherein C _30 and R _30 comprise 30 tags, C _50 and R _50 comprise 50 tags, C _100 and R _100 comprise 100 tags, and C _500 and R _500 comprise 500 tags. The label distribution mode in C _30, C _50 and C _100 is cluster distribution; the labels in R _30, R _50 and R _100 are distributed uniformly; the labels in C _500 and R _500 are generated by linear combination of the labels in C _100 and R _100, respectively. The threshold range of the power radius of the reader is T ═ epsilonmin,εmax]In which epsilonmin＝8，εmax15. The spatial boundaries of the C _30, C _50, C _100, R _30, R _50, and R _100 test samples were 50m, and the spatial boundaries of the C _500 and R _500 test samples were 150 m. The population size N is set to 10, and the optimized generation gen is set to 1000.
3. Simulation content and result analysis thereof:
simulation 1, different data sets were simulated with the present invention, and the results are shown in fig. 5, where:
FIG. 5(a) is a diagram illustrating the results of a simulation performed on data set C _ 30;
FIG. 5(b) is a diagram illustrating the results of a simulation performed on data set C _ 50;
FIG. 5(C) is a diagram showing the results of a simulation performed on the data set C _ 100;
FIG. 5(d) is a diagram illustrating the results of a simulation performed on the data set R _ 30;
FIG. 5(e) is a diagram illustrating the results of a simulation performed on the data set R _ 50;
FIG. 5(f) is a diagram illustrating the results of a simulation performed on the data set R _ 100;
FIG. 5(g) is a diagram illustrating the results of a simulation performed on data set C _ 500;
fig. 5(h) is a diagram illustrating the result of simulation of the data set R _ 500.
As can be seen from fig. 5, the requirements of the RFID network planning can be effectively met for the test samples with different numbers of tags and the test samples distributed with different tags, the number of the readers required, the size of the power radius of the readers and the positions of the readers in the working area are reasonably determined, and the RFID network planning is effectively achieved.
The performance values of the results of the RFID network planning in the 6 test samples C _30, C _50, C _100, R _30, R _50, and R _100 in fig. 5 are counted, and the results are shown in table 1.
Table 1 summary of sample test results
Figure BDA0002359657460000111
As can be seen from table 1, after the RFID network is planned by the present invention, the coverage of the system in 6 test examples, C _30, C _50, C _100, R _30, R _50, and R _100, can reach 100%, the inter-reader interference rate can be reduced to 0, and the total power and the load balance of the RFID system obtained by the present invention are both numerically smaller than the total power and the load balance of the RFID system obtained by the existing curling algorithm and the MOEAD algorithm, which indicates that the present invention can effectively meet the requirements of the RFID network planning.
And 2, simulating the result of the coverage rate obtained after the RFID network planning and changing along with the number of the readers by using the method, and comparing the result with the existing curling algorithm and the MOEAD algorithm, wherein the result is shown in figure 6.
As can be seen from FIG. 6, compared with the existing curling algorithm and MOEAD algorithm, the method of the invention can effectively improve the coverage rate of the RFID network and realize the full coverage of the RFID network.
And 3, simulating the result of the change of the interference rate obtained after the RFID network planning along with the number of the readers by using the method, and comparing the result with the existing curling algorithm and the MOEAD algorithm, as shown in fig. 7.
As can be seen from FIG. 7, compared with the existing curling algorithm and MOEAD algorithm, the method of the invention can effectively reduce the coverage rate between readers in the RFID network and ensure the communication between the readers and the tags in the RFID network.
It should be noted that: the embodiment of performing the RFID network planning by using the RFID network planning method based on the hybrid particle swarm algorithm is only used as an illustration in the test sample, and is not used to limit the present invention, and the method can be applied to other scenes according to actual needs. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present invention are only embodiments of the present invention and shall be included in the protection scope of the present invention.

Claims (7)

1. A RFID network planning method based on a hybrid particle swarm algorithm is characterized by comprising the following steps:
(1) simulating particle swarm motion to construct a wireless radio frequency network RFID system model, and respectively defining the distribution of a network working area, a tag and a reader;
(2) defining the initial value of the number of readers in the network as: n is 3;
(3) classifying the labels in the network by using a K mean value algorithm to obtain m readers and a label classification result TS of each readerjWhere j is ∈ [0, m ]](ii) a The method is realized as follows:
(3a) randomly initializing position coordinates of m readers
Figure FDA0003173723690000011
(3b) Assigning the labels to the clusters closest to the readers by taking each reader as a cluster center;
(3c) calculating the mean value of the label coordinates in each cluster to update the cluster center:
Figure FDA0003173723690000012
Figure FDA0003173723690000013
wherein the content of the first and second substances,
Figure FDA0003173723690000014
denotes the updated coordinates of the jth reader, (x)ij,yij) Indicates the ith distribution in the reader rjPosition coordinates of the label within the coverage, NtjIs shown in the reader rjThe number of tags within the coverage of (c);
(3d) cluster center coordinates to be updated
Figure FDA0003173723690000015
From the original cluster center coordinates
Figure FDA0003173723690000016
And (3) comparison:
if the updated cluster center coordinates
Figure FDA0003173723690000017
From the original cluster center coordinates
Figure FDA0003173723690000018
If the two are equal, the classification is finished;
if the updated cluster center coordinates
Figure FDA0003173723690000019
From the original cluster center coordinates
Figure FDA00031737236900000110
If not, returning to the step (3 b);
(4) calculating the power radius u of each readerjAnd setting a threshold range T ═ epsilon of the power radius according to the requirements in practical engineeringmin,εmax]In which epsilonminIs the minimum value of the available reader power radius, epsilonmaxThe maximum value of the available reader power radius;
(5) judgment ujWhether threshold range T is satisfied:
if ujWithin the threshold value range T, determining the number of readers in the network as m;
if ujIf the number m of the readers in the network is not within the threshold range T, the number m of the readers in the network is adjusted to be m +1, and the step (3) of reclassifying is returned;
(6) according to the number of readers in a network, randomly obtaining the initial position and the power radius of each reader through a K-means algorithm, taking the positions and the power radii of all the readers as a particle, and repeatedly executing N times to obtain N particles as a population;
(7) evaluating each performance of the particles in the population by using four fitness functions, namely a coverage rate COV, an interference rate INT, a total power POW and a load balance degree LB in sequence;
(8) recording the optimal value of each particle, namely the individual optimal particle, and the optimal value of all particles in the population, namely the global optimal particle, in a layered mode according to the evaluation result of the step (7); the method is realized as follows:
(6a) judging whether the population is the first generation:
if the population is the first generation population, the individual optimal particle is the individual optimal particle;
if not, executing (6 b);
(6b) respectively comparing the coverage rate of the particles in the current population with the coverage rate of the individual optimal particles:
if the two are equal, executing (6 c);
if the coverage rate of the particles in the current population is larger than that of the individual optimal particles, executing (6 f);
if the coverage rate of the particles in the current population is smaller than that of the individual optimal particles, ending;
(6c) respectively comparing the interference rates of the particles in the current population and the individual optimal particles:
if the interference rate of the particles in the current population is equal to the interference rate of the individual optimal particles, executing (6 d);
if the interference rate of the particles in the current population is smaller than the interference rate of the individual optimal particles, executing (6 f);
if the interference rate of the particles in the current population is greater than the interference rate of the individual optimal particles, ending;
(6d) respectively comparing the total system power of the particles in the current population with the optimal individual particles:
if the total system power of the particles in the current population is equal to the total system power of the individual optimal particles, executing (6 e);
if the total system power of the particles in the current population is less than that of the individual optimal particles, executing (6 f);
if the total system power of the particles in the current population is greater than that of the individual optimal particles, ending;
(6e) respectively comparing the load balance degrees of the particles in the current population and the individual optimal particles:
if the load balance degree of the particles in the current population is smaller than the total system power of the individual optimal particles, executing (6 f);
if the load balance degree of the particles in the current population is more than or equal to the total system power of the individual optimal particles, ending;
(6f) updating the current particle as an individual optimal particle, and performing hierarchical comparison on the four fitness values of the individual optimal particle and the global optimal particle respectively to update the individual optimal particle as the global optimal particle;
(9) setting the iteration times of the population as gen, and updating the speed and the position of the particles in an iterative manner by utilizing a particle swarm algorithm updating strategy according to the individual optimal particles and the global optimal particles obtained in the step (8);
(10) and adjusting the position of the reader in each particle according to the covering state of the label:
if the label is covered by a plurality of readers at the same time, repeatedly selecting any one of the two readers with the nearest distance for position adjustment, so that the coordinate position changes after mutual repulsion between the readers until the label is covered by only one reader;
if the label is not covered by any reader, the reader closest to the label is moved to the position of the label, and then the coordinate is changed until the label is covered by the reader;
(11) and (5) repeatedly executing the steps (7) to (10) until the maximum population algebra gen is met, and obtaining the global optimal particles, wherein the optimal particles are the RFID network planning scheme, namely the deployment position and the power of the reader.
2. The method of claim 1, wherein simulating the motion of the group of particles in (1) constructs a model of a wireless radio frequency network (RFID) system by:
(2a) defining an RFID network working area to be optimized as a particle swarm motion space, wherein the space boundary is 50m multiplied by 50 m;
(2b) label t to be coverediRandomly distributed in the working space in a uniform or cluster form, and the coordinates are expressed as
Figure FDA0003173723690000031
(2c) Reader rjIs expressed as
Figure FDA0003173723690000032
3. The method of claim 1, wherein the power radius of each reader is calculated in (4) by the following equation:
uj＝max(d(tij,rj))tij∈TSj
Figure FDA0003173723690000033
wherein r isjJ-th reader, u, representing a particlejIndicates the power radius of the jth reader,
Figure FDA0003173723690000034
indicating the position coordinates of the jth reader; t is tijIs shown in the reader rjThe ith tag in the coverage of (1),
Figure FDA0003173723690000035
indicating the location coordinates of the tag; d (t)ij,rj) Presentation reader rjAnd a label tijEuclidean distance of.
4. The method according to claim 1, wherein the evaluation values of the performance of the RFID network by using different fitness functions in the step (7) are calculated as follows:
label coverage rate:
Figure FDA0003173723690000036
wherein the content of the first and second substances,
Figure FDA0003173723690000037
ujdenotes the power radius of the jth reader, p (t)i,rj) Indicates the label tiIs read by the reader rjOver-coverage state of, NtIs the number of tags in the network, l (t)i,rj) Is a reader rjAnd a label tiThe Euclidean distance of;
inter-reader interference rate:
Figure FDA0003173723690000041
wherein
Figure FDA0003173723690000042
num(ti) To cover the label tiThe number of readers, rat (t)i) Is a label tiThe degree of interference of (c);
total power of the RFID system:
Figure FDA0003173723690000043
wherein RS represents a set of readers in the network;
and (3) load balance degree of the RFID system:
Figure FDA0003173723690000044
wherein the content of the first and second substances,
Figure FDA0003173723690000045
is a reader rjThe number of labels covered.
5. The method of claim 1, wherein the positions of the particles are updated in (9) by a particle swarm algorithm, by the following formula:
Figure FDA0003173723690000046
wherein the content of the first and second substances,
Figure FDA0003173723690000047
indicates the position of the ith particle in the population of the (d + 1) th generation,
Figure FDA0003173723690000048
denotes the position, V, of the ith particle in the population of the d generationi d+1Represents the velocity of the ith particle in the population of the (d + 1) th generation;
Figure FDA0003173723690000049
Vi drepresenting the velocity of the ith particle in the current generation-d population,
Figure FDA00031737236900000410
represents the optimal position of the ith particle in the population of the d generation, gdRepresents the optimal position of all particles in the population of the d generation, c1，c2Representing two different learning factorsC is prepared from1＝1，c2＝2，r1 d
Figure FDA00031737236900000411
Is between [0,1]And a random number different from the above, wherein omega represents an inertia factor, and omega is 0.5.
6. The method of claim 1, wherein the coordinate position changes after mutual repulsion between readers in the system (10), and the calculation formula is as follows:
Figure FDA00031737236900000412
Figure FDA00031737236900000413
wherein the content of the first and second substances,
Figure FDA00031737236900000414
are respectively the first reader r1And a second reader r2A is between [0,1 ]]A random number in between, and a random number,
Figure FDA00031737236900000415
is a second reader r2Coordinates obtained after rejection.
7. The method of claim 1, wherein the reader in the reader (10) is moved closer to the tag position, so that the reader coordinates are changed, and the calculation formula is as follows:
Figure FDA0003173723690000051
Figure FDA0003173723690000052
wherein the content of the first and second substances,
Figure FDA0003173723690000053
are respectively the first reader r1And a second reader r2Beta is between [0,1 ]]A random number in between, and a random number,
Figure FDA0003173723690000054
is a reader r2And (4) coordinates obtained after the label position is closed.
CN202010018027.9A 2020-01-08 2020-01-08 RFID network planning method based on hybrid particle swarm optimization Active CN111225367B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010018027.9A CN111225367B (en) 2020-01-08 2020-01-08 RFID network planning method based on hybrid particle swarm optimization

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010018027.9A CN111225367B (en) 2020-01-08 2020-01-08 RFID network planning method based on hybrid particle swarm optimization

Publications (2)

Publication Number Publication Date
CN111225367A CN111225367A (en) 2020-06-02
CN111225367B true CN111225367B (en) 2022-03-04

Family

ID=70826122

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010018027.9A Active CN111225367B (en) 2020-01-08 2020-01-08 RFID network planning method based on hybrid particle swarm optimization

Country Status (1)

Country Link
CN (1) CN111225367B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112004233B (en) * 2020-08-31 2022-08-02 南京邮电大学 Network planning method based on big data mining
CN113111594B (en) * 2021-05-12 2022-04-12 中国人民解放军国防科技大学 Multi-objective optimization-based frequency planning method and device and computer equipment

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103793551A (en) * 2013-12-02 2014-05-14 电子科技大学 Deployment algorithm of large-scale RFID (radiofrequency identification) readers in three-dimensional space
CN104504424A (en) * 2014-12-27 2015-04-08 西安电子科技大学 Multi-population symbiotic evolution based radio frequency identification network layout optimization method
CN106372678A (en) * 2015-07-23 2017-02-01 中国科学院沈阳自动化研究所 RFID network reader scheduling optimization method based on multi-swarm particle swarm algorithm

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8072311B2 (en) * 2008-04-14 2011-12-06 Mojix, Inc. Radio frequency identification tag location estimation and tracking system and method
US9443117B2 (en) * 2012-12-14 2016-09-13 Symbol Technologies, Llc Self-optimizing method of and system for efficiently deploying radio frequency identification (RFID) tag readers in a controlled area containing RFID-tagged items to be monitored
CN103310248A (en) * 2013-04-25 2013-09-18 中山大学 Optimized RFID (Radio Frequency Identification) network system based on particle swarm algorithm
CN104517141B (en) * 2014-12-27 2017-06-13 西安电子科技大学 Radio frequency identification network topology method based on load balance Yu particle cluster algorithm
CN105956504B (en) * 2016-04-20 2018-03-06 西安电子科技大学 The RFID network planing method of simulated curling motion

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103793551A (en) * 2013-12-02 2014-05-14 电子科技大学 Deployment algorithm of large-scale RFID (radiofrequency identification) readers in three-dimensional space
CN104504424A (en) * 2014-12-27 2015-04-08 西安电子科技大学 Multi-population symbiotic evolution based radio frequency identification network layout optimization method
CN106372678A (en) * 2015-07-23 2017-02-01 中国科学院沈阳自动化研究所 RFID network reader scheduling optimization method based on multi-swarm particle swarm algorithm

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于分数阶达尔文粒子群的不等间距节点部署优化算法;张利峰;《计算机应用与软件》;20190512(第05期);全文 *

Also Published As

Publication number Publication date
CN111225367A (en) 2020-06-02

Similar Documents

Publication Publication Date Title
Yin A discrete particle swarm algorithm for optimal polygonal approximation of digital curves
Yin Multilevel minimum cross entropy threshold selection based on particle swarm optimization
Liu et al. A new data classification method based on chaotic particle swarm optimization and least square-support vector machine
CN109902953B (en) Power consumer classification method based on self-adaptive particle swarm clustering
CN111225367B (en) RFID network planning method based on hybrid particle swarm optimization
CN107203785A (en) Multipath Gaussian kernel Fuzzy c-Means Clustering Algorithm
CN106971091B (en) Tumor identification method based on deterministic particle swarm optimization and support vector machine
CN109814066B (en) RSSI indoor positioning distance measurement method and indoor positioning platform based on neural network learning
CN104850867A (en) Object identification method based on intuitive fuzzy c-means clustering
CN109299698A (en) A kind of wireless channel scene recognition method based on support vector machines
CN103456017B (en) Image partition method based on the semi-supervised weight Kernel fuzzy clustering of subset
US11315283B1 (en) Method and system for object detection and posture estimation based on object-customized image feature detection algorithm
CN113434856A (en) Network intrusion detection method based on PSOGWO-SVM algorithm
CN113255873A (en) Clustering longicorn herd optimization method, system, computer equipment and storage medium
Zhao et al. Improved grey wolf optimization based on the two-stage search of hybrid CMA-ES
Javani et al. Clustering and feature selection via PSO algorithm
CN105512675A (en) Memory multi-point crossover gravitational search-based feature selection method
Wang et al. Research on classification and recognition of attacking factors based on radial basis function neural network
Guo et al. Active semi-supervised K-means clustering based on silhouette coefficient
CN111639688B (en) Local interpretation method of Internet of things intelligent model based on linear kernel SVM
Garbuglia et al. Bayesian optimization for microwave devices using deep GP spectral surrogate models
Zhang et al. An improved least squares SVM with adaptive PSO for the prediction of coal spontaneous combustion
CN105956504B (en) The RFID network planing method of simulated curling motion
Yao et al. Power function-based signal recovery transition optimization model of emergency traffic
CN115393631A (en) Hyperspectral image classification method based on Bayesian layer graph convolution neural network

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