CN105960009A - Positioning method in software defined wireless sensor network - Google Patents

Abstract

The invention provides a positioning method in a software defined wireless sensor network. The method comprises the steps of setting a contribution matrix for nodes in a network, wherein a certain element in the matrix is the contribution value of an anchor node in the row which corresponds with the element to a blind node positioning result in the line that corresponds with the element; according to the set matrix, constructing an optimization problem by means of knowability of a controller in the software defined network to global network information, selecting a positioning node from all anchor nodes for the blind node through a 0-1 programming method, wherein a requirement that the sum of powers of selected links is not larger than the total power of the wireless sensor network is satisfied, and maximizing the contribution value of the selected links in the network on condition that the requirement is satisfied; and calculating the position of the blind node by means of a linear least square algorithm by means of the known position information of the selected positioning node. The positioning method can be used for selecting a node which most facilitates positioning on condition that the power of the wireless sensor network is limited and furthermore improves positioning precision.

Description

A kind of localization method in software definition wireless sensor network

Technical field

The present invention relates to wireless location technology field, the location side in a kind of software definition wireless sensor network Method.

Background technology

In many wireless communication systems, it is possible to know that the positional information of nodes has become as an actual demand, Obtain wide variety of wireless sensor network (Wireless SensorNetworks, WSN) especially at present.Due to In some specific localizing environment, often cannot accurately receive GPS (Global Positioning System) satellite letter Number, thus the application of whole network is had extremely critical by the localization method that research the most accurately calculates sensor node position Effect.

Consider the feature of wireless sensor network power limited, how in the case of meeting network total power requirements, to study height The localization method of precision is most important.Owing to lacking the knowability to the whole network information, at present at wireless sensor network The effective power distribution localization method of lower research is distributed mostly, has certain journey relative to the distribution method of overall situation formula Performance inferior position on degree.

Summary of the invention

Goal of the invention: for solving above-mentioned technical problem, the present invention, on the basis of software defined network technology, utilizes software The definition network controller knowability to global network information, proposes the location in a kind of software definition wireless sensor network Method.The method can improve node locating performance in the case of meeting wireless sensor network power requirement, has wide Application prospect.

Technical scheme: for realizing above-mentioned technique effect, the technical scheme is that a kind of software definition wireless senser Localization method in network, described software definition wireless sensor network includes software defined network controller, N_aIndividual blind Node and N_bAnchor node；The method comprising the steps of:

(1) any one anchor node is calculated in described software definition wireless sensor network to any one blind node locating Result contribution margin, builds a contribution matrix with all contribution margins calculatedIn contribution matrix Element w_ijThe expression anchor node j positioning result contribution margin to blind node i, and i ∈ [1,2 ..., N_a], j ∈ [1,2 ..., N_b]；

(2) location node selection matrix is builtDefinition f_ijAnchor node j conduct is chosen in=1 expression Blind node i be set to node, f_ij=0 represents that anchor node j is set to node not as blind node i；

(3) to maximize the contribution matrix product with location node selection matrix transpose as target problem, it may be assumed that

${\max }_F{\mathrm{\Σ}}_{j=N_a+1}^{N_a+N_b}{\mathrm{\Σ}}_{i=1}^{N_a}(Q\·F^T)$

Arranging constraints is:

①f_ij=1 or f_ij=0；

②

③

Wherein, ε_ijRepresent the through-put power on the link that anchor node j and blind node i are constituted,Represent network Big general power；Solve the best located node selection matrix simultaneously meeting above three constraints；

(4) software defined network controller is that each blind node selection positions node according to best located node selection matrix, And according to the positional information of selected location node, use linear least-squares algorithm to calculate the position of blind node.

Further, described step (1) builds contribution matrixMethod be:

(1-1) obtain blind node i and receive the signal intensity P of anchor node j_ijFor:

P_ij=P₀-10αlog₁₀d_ij+v

Wherein, P₀For the received signal strength at distance anchor node j 1m place, unit be dBm, α be described software calmly The path attenuation factor in justice wireless sensor network, v is the stochastic variable of Gaussian distributed, And according to signal intensity P_ijCalculate distance d between blind node i and anchor node j_ij；

(1-2) according to the d obtained in step (2)_ijCalculating anchor node j to the positioning result contribution margin of blind node i is:

$w_{ij}=\frac{1}{{\mathrm{\σ}}_{ij}^2}$

${\mathrm{\σ}}_{ij}^2=\frac{1}{b}{\mathrm{\Σ}}_{k\∈\left\{j\right\},k\≠j}\frac{{\mathrm{\Σ}}_{k=2}^{N_a+N_b}{d}_{jk}^{-2}}{{\mathrm{\Σ}}_{k=N_a+1}^{N_a+N_b-1}\left(\frac{d_{i\⊥jk}{d}_{jk}}{d_{ij}^2{d}_{ik}^2}\right)}$

Wherein,d_ijAnd d_ikRepresent blind node i and anchor node j and the distance of anchor node k, anchor respectively Node k is any one anchor node being different from anchor node j；d_jkRepresent the distance between anchor node j and anchor node k； d_i⊥jkRepresent that blind node i is to anchor node j and the beeline of the line of anchor node k；

(1-3) based on all w calculated in step (1-2)_ijBuild and contribute matrix Q:

$Q={\[w_{ij}\]}_{N_a\×N_b}.$

Further, technique scheme further comprises the steps of: before carrying out described step (3), initialize location Node selection matrixThe method initializing location node selection matrix is:

Distance d between any blind node i and the arbitrarily anchor node j that obtain according to step (1-1)_ij, it is judged that anchor node j Whether in the communication range of blind node i；If the determination result is YES, then f is made_ij=1, otherwise make f_ij=0.

Further, the method for the position calculating blind node in described step (4) is:

(4-1) the location joint that software defined network controller is chosen as blind node i is set according to best located node selection matrix Point number be N, define blind node i receive location node l signal strength values be { P_il, l ∈ [1,2 ..., N]；

(4-2) calculating the distance estimations vector of blind node i according to step (1-1) is:

$\hat{d}=\[{\hat{d}}_{i1},{\hat{d}}_{i2},...,{\hat{d}}_{iN}\]$

WhereinRepresent the estimated distance value of location node l and blind node i,

(4-3) from the node of N number of location, the minimum node r of estimated distance value is chosen as reference mode, i.e.

$r=\arg {\min }_l\left\{{\hat{d}}_{il}\right\}$

Utilize distance estimations vector, structure equation below:

$A_{lls}x=\frac{1}{2}{p}_{lls}$

$A_{lls}=\left[\begin{array}{cc}{x}_1-x_r& y_1-y_r\\ x_2-x_r& y_2-y_r\\ \·& \·\\ \·& \·\\ \·& \·\\ x_N-x_r& y_N-y_r\end{array}\right]$

$p_{lls}=\left[\begin{array}{c}{\hat{d}}_{ri}^2-{\hat{d}}_{1i}^2-k_{r1}\\ {\hat{d}}_{ri}^2-{\hat{d}}_{2i}^2-k_{r2}\\ \·\\ \·\\ \·\\ {\hat{d}}_{ri}^2-{\hat{d}}_{Ni}^2-k_{rN}\end{array}\right]$

$k_{il}=x_r^2+y_r^2-x_l^2-y_l^2,l\≠r$

Wherein, x={ (x_i, y_i)∈R²Represent that the location estimation of blind node i is vectorial；

According to LLS algorithm, solving equationThe location estimation vector obtaining blind node i is:

$x=\frac{1}{2}{\left(A_{lls}^T{A}_{lls}\right)}^{-1}{A}_{lls}^T{p}_{lls}.$

Beneficial effect: compared with prior art, present invention have the advantage that

The present invention utilizes the software defined network controller knowability to global network information, by building a global optimum Change problem, is converted into an one-zero programming problem by positioning node problems for the blind node selection in network, thus reaches Meet the condition of the general power that all power sum no more than networks choosing link can bear in wireless sensor network Under, for each blind node selection most beneficial for the location node of its positioning result so that the contribution to location of the selected link Value sum reaches maximum, improves positioning precision.

Accompanying drawing explanation

Fig. 1 is the flow chart of the embodiment of the present invention；

Fig. 2 is the location scene graph of software definition wireless sensor network.

Detailed description of the invention

For making the purpose of the embodiment of the present invention, technical scheme and advantage clearer, below in conjunction with in the embodiment of the present invention Accompanying drawing, the technical scheme in the embodiment of the present invention is carried out clear, complete description, it is clear that described embodiment It is a part of embodiment of the present invention rather than whole embodiments.

Embodiment: in software definition wireless sensor network, it is assumed that have N in network_aThe blind joint that individual coordinate position is unknown Point, and N_bAnchor node known to individual coordinate position.As shown in Figure 2, in this software definition wireless sensor network, The information of global network grasped by controller, chooses, for the distribution of blind node, the best located node obtained from all anchor nodes.

Accompanying drawing 1 describes the process that in this network, blind node location is estimated, in this wireless sensor network, blind node is to anchor The received signal strength value (Received Signal Strength Indicator, RSSI) of node uses following model:

P_d=P₀-10α log₁₀d+v (1)

In formula (1), P₀For the received signal strength at distance anchor node 1m, unit is dBm, P_dIt is should in distance The RSSI, α of the blind node at anchor node d is the path attenuation factor set according to wireless propagation environment, and v is to obey The stochastic variable of Gauss distribution,

Assume that in network, a certain blind node i receives the signal intensity P of anchor node j_ij(dBm) obey following Gauss to divide Cloth:

$P_{ij}\left(dBm\right)~N\left({\stackrel{\‾}{P}}_{ij}\right(dBm),{\mathrm{\σ}}_{dB}^2)---\left(2\right)$

Calculate the following expression of any two anchor node m and n in blind node i and network:

${\mathrm{\σ}}_{i\_mn}^2=\frac{1}{b}\frac{{\mathrm{\Σ}}_{m=N_a+1}^{N_a+N_b}{d}_{im}^{-2}}{{\mathrm{\Σ}}_{m=N_a+1}^{N_a+N_b-1}{\mathrm{\Σ}}_{n=m+1}^{N_a+N_b}{\left(\frac{d_{i\⊥mn}{d}_{mn}}{d_{im}^2{d}_{in}^2}\right)}^2}---\left(3\right)$

In formula (3),d_im、d_inAnd d_mnRepresent blind node i and anchor node m's and n respectively Distance between distance and anchor node m and n, d_i⊥mnIt it is the blind node i short distance to the line of anchor node m and n From.

For a certain given blind node i, in the network, arbitrary anchor node m and other N_b-1 anchor node combination energy Obtain N_bThe transition formula evaluation of-1 such as formula (3), by this N_b-1 value summation obtains

${\mathrm{\σ}}_{im}^2=\frac{1}{b}{\mathrm{\Σ}}_{b_k\∈\left\{b_j\right\}\backslash b_m}\frac{{\mathrm{\Σ}}_{k=2}^{N_a+N_b}{d}_{mk}^{-2}}{{\mathrm{\Σ}}_{k=N_a+1}^{N_a+N_b-1}{\left(\frac{d_{i\⊥mk}{d}_{mk}}{d_{im}^2{d}_{ik}^2}\right)}^2}---\left(4\right)$

The inverse of this and value is defined as the anchor node m contribution margin to blind node i positioning result, i.e.

$w_{im}=\frac{1}{{\mathrm{\σ}}_{im}^2}---\left(5\right)$

Build and contribute matrix as follows:

In formula (6), element w_im(i=1,2 ..., N_a, m=N_a+ 1, N_a+ 2 ..._Na+N_b) represent anchor The node m contribution margin to blind node i positioning result.

Build location as follows node selection matrix:

Element f in this matrix_ij(i=1,2 ..., N_a, m=N_a+ 1, N_a+ 2 ... N_a+N_b) value be 0 Or 1, f_ijThe location node that anchor node j positions is chosen in=1 expression as blind node i, if f_ij=0, then anchor node The location node that j positions not as blind node i.

Before location starts, initialize location node selection matrix, if anchor node j is in the communication range of blind node i, Then f_ij=1, otherwise f_ij=0.

Consider the problem such as wireless sensor network power limited, build following global optimization problem:

${\max }_F{\mathrm{\Σ}}_{j=N_a+1}^{N_a+N_b}{\mathrm{\Σ}}_{i=1}^{N_a}(Q\·F^T)---\left(8\right)$

Constraints:

f_ij=1 or f_ij=0 (9)

${\mathrm{\Σ}}_{j=N_a+1}^{N_a+N_b}{\mathrm{\Σ}}_{i=1}^{N_a}{\mathrm{\ϵ}}_{ij}\≤P_{\lim }^{\left(tot\right)}---\left(10\right)$

${\mathrm{\Σ}}_{j=N_a+1}^{N_a+N_b}{f}_{ij}\≥3---\left(11\right)$

In this optimization problem, object function represents that solving best located node selection matrix makes the location of whole network Contribution margin sum reaches maximum.Constraints (9) represents that the element value in the node selection matrix of location is 0 or 1；(10) Middle ε_ijRepresent the through-put power on the link that anchor node j and blind node i are constituted,Represent the maximum general power of network, The power sum of all links participating in location not can exceed that this value；Constraints (11) represents in two-dimensional localization, root According to triangle characteristic, the location estimation of each blind node at least needs 3 location nodes.

According to solving the best located node selection matrix that this optimization problem obtains, software defined network controller is blind joint Point is chosen for the location node of its location.Be assumed to be blind node i choose N number of anchor node as location node, receive Signal strength values is designated as { P_il, l=1,2 ..., N}, according to formula (1) calculate the estimation of each location node and blind node i away from Distance values is

${\hat{d}}_{il}={10}^{\frac{P_0-P_{il}}{10\mathrm{\α}}}---\left(12\right)$

N number of range estimation constitutes distance estimations vectorUnderneath with a linear young waiter in a wineshop or an inn Multiplication algorithm calculates the position of blind node i.The node r that chosen distance estimated value is minimum from the node of N number of location is as linearly The reference mode of least-squares algorithm:

$r=\arg {\min }_l\left\{{\hat{d}}_{il}\right\},l=1,2,...,N---\left(13\right)$

Utilize distance estimations vector, structure equation below:

$A_{lls}x=\frac{1}{2}{p}_{lls}---\left(14\right)$

In formula (14), x={ (x_i, y_i)∈R²Represent blind node location estimate vector, and by following expression:

$A_{lls}=\left[\begin{array}{cc}{x}_1-x_r& y_1-y_r\\ x_2-x_r& y_2-y_r\\ \·& \·\\ \·& \·\\ \·& \·\\ x_N-x_r& y_N-y_r\end{array}\right],p_{lls}=\left[\begin{array}{c}{\hat{d}}_{ri}^2-{\hat{d}}_{1i}^2-k_{r1}\\ {\hat{d}}_{ri}^2-{\hat{d}}_{2i}^2-k_{r2}\\ \·\\ \·\\ \·\\ {\hat{d}}_{ri}^2-{\hat{d}}_{Ni}^2-k_{rN}\end{array}\right]---\left(15\right)$

$k_{il}=x_r^2+y_r^2-x_l^2-y_l^2,l\≠r---\left(16\right)$

According to LLS algorithm, solve and obtain the position of blind node and be

$x=\frac{1}{2}{\left(A_{lls}^T{A}_{lls}\right)}^{-1}{A}_{lls}^T{p}_{lls}---\left(17\right).$

The above is only the preferred embodiment of the present invention, it should be pointed out that: for those skilled in the art For, under the premise without departing from the principles of the invention, it is also possible to make some improvements and modifications, these improvements and modifications are also Should be regarded as protection scope of the present invention.

Claims (4)

1. the localization method in a software definition wireless sensor network, it is characterised in that described software definition is wireless Sensor network includes software defined network controller, N_aIndividual blind node and N_bAnchor node；

The method comprising the steps of:

(1) any one anchor node is calculated in described software definition wireless sensor network to any one blind node locating Result contribution margin, builds a contribution matrix with all contribution margins calculatedIn contribution matrix Element w_ijThe expression anchor node j positioning result contribution margin to blind node i, and i ∈ [1,2 ..., N_a], j ∈ [1,2 ..., N_b]；

(2) location node selection matrix is builtDefinition f_ijAnchor node j conduct is chosen in=1 expression Blind node i be set to node, f_ij=0 represents that anchor node j is set to node not as blind node i；

(3) to maximize the contribution matrix product with location node selection matrix transpose as target problem, it may be assumed that

${\max }_F{\mathrm{\Σ}}_{j=N_a+1}^{N_a+N_b}{\mathrm{\Σ}}_{i=1}^{N_a}(Q\·F^T)$

Arranging constraints is:

①f_ij=1 or f_ij=0；

②

③

Wherein, ε_ijRepresent the through-put power on the link that anchor node j and blind node i are constituted,Represent network Big general power；Solve the best located node selection matrix simultaneously meeting above three constraints；

(4) software defined network controller is that each blind node selection positions node according to best located node selection matrix, And according to the positional information of selected location node, use linear least-squares algorithm to calculate the position of blind node.

Localization method in a kind of software definition wireless sensor network the most according to claim 1, its feature exists In, described step (1) builds contribution matrixMethod be:

(1-1) obtain blind node i and receive the signal intensity P of anchor node j_ijFor:

P_ij=P₀-10αlog₁₀d_ij+v

Wherein, P₀For the received signal strength at distance anchor node j1m place, unit be dBm, α be described software calmly The path attenuation factor in justice wireless sensor network, υ is the stochastic variable of Gaussian distributed, Variance for Gauss distribution；And according to signal intensity P_ijCalculate distance d between blind node i and anchor node j_ij；

(1-2) according to the d obtained in step (2)_ijCalculating anchor node j to the positioning result contribution margin of blind node i is:

$w_{ij}=\frac{1}{{\mathrm{\σ}}_{ij}^2}$

${\mathrm{\σ}}_{ij}^2=\frac{1}{b}{\mathrm{\Σ}}_{k\∈\left\{j\right\},k\≠j}\frac{{\mathrm{\Σ}}_{k=2}^{N_a+N_b}{d}_{jk}^{-2}}{{\mathrm{\Σ}}_{k=N_a+1}^{N_a+N_b-1}\left(\frac{d_{i\⊥jk}{d}_{jk}}{d_{ij}^2{d}_{ik}^2}\right)}$

Wherein,d_ijAnd d_ikRepresent blind node i and anchor node j and the distance of anchor node k, anchor respectively Node k is any one anchor node being different from anchor node j；d_jkRepresent the distance between anchor node j and anchor node k； d_i⊥jkRepresent that blind node i is to anchor node j and the beeline of the line of anchor node k；

(1-3) based on all w calculated in step (1-2)_ijBuild and contribute matrix Q:

$Q={\[w_{ij}\]}_{N_a\×N_b}.$

Localization method in a kind of software definition wireless sensor network the most according to claim 2, its feature exists In, further comprise the steps of: before carrying out described step (3), initialize location node selection matrix The method initializing location node selection matrix is:

Distance d between any blind node i and the arbitrarily anchor node j that obtain according to step (1-1)_ij, it is judged that anchor node j Whether in the communication range of blind node i；If the determination result is YES, then f is made_ij=1, otherwise make f_ij=0.

Localization method in a kind of software definition wireless sensor network the most according to claim 2, its feature exists In, the method for the position calculating blind node in described step (4) is:

(4-1) the location joint that software defined network controller is chosen as blind node i is set according to best located node selection matrix Point number be N, define blind node i receive location node l signal strength values be { P_il, l ∈ [1,2 ..., N]；

(4-2) calculating the distance estimations vector of blind node i according to step (1-1) is:

$\hat{d}=\[{\hat{d}}_{i1},{\hat{d}}_{i2},...,{\hat{d}}_{iN}\]$

WhereinRepresent the estimated distance value of location node l and blind node i,

(4-3) from the node of N number of location, the minimum node r of estimated distance value is chosen as reference mode, i.e.

$r=\arg {\min }_l\left\{{\hat{d}}_{il}\right\}$

Utilize distance estimations vector, structure equation below:

$A_{lls}x=\frac{1}{2}{p}_{lls}$

$A_{lls}=\left[\begin{array}{cc}{x}_1-x_r& y_1-y_r\\ x_2-x_r& y_2-y_r\\ \begin{array}{c}.\\ .\\ .\end{array}& \begin{array}{c}.\\ .\\ .\end{array}\\ x_N-x_r& y_N-y_r\end{array}\right]$

$p_{lls}=\left[\begin{array}{c}{\hat{d}}_{ri}^2-{\hat{d}}_{1i}^2-k_{r1}\\ {\hat{d}}_{ri}^2-{\hat{d}}_{2i}^2-k_{r2}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ {\hat{d}}_{ri}^2-{\hat{d}}_{Ni}^2-k_{rN}\end{array}\right]$

$k_{il}=x_r^2+y_r^2-x_l^2-y_l^2,l\≠r$

Wherein, x={ (x_i, y_i)∈R²Represent that the location estimation of blind node i is vectorial；

According to LLS algorithm, solving equationThe location estimation vector obtaining blind node i is:

$x=\frac{1}{2}{\left(A_{lls}^T{A}_{lls}\right)}^{-1}{A}_{lls}^T{p}_{lls}.$