CN101938832A - Division and refinement-based node self-positioning method for wireless sensor network - Google Patents

Abstract

The invention discloses a division and refinement-based node self-positioning method for a wireless sensor network, belonging to the intersection field of embedded development and wireless communication. The method comprises the following steps: converting received signal strength of anchor nodes into distance, and ranking the anchor nodes according to positions; and selecting two anchor nodes per time according to the sequence to form a positioning combination; initially estimating the coordinates of an unknown node according to the position of the anchor node and corresponding ranging information; verifying a coordinate estimated valve by selecting effective retardant anchor node information; calculating the weight corresponding to each estimated value; and outputting a positioning result by weighted refinement. The method does not require additional hardware or increase communication quantity, and is easy to implement. Simulation and comparison are carried out on the method of the invention with a centroid method and a weighted centroid method through MATLAB, and a result proves that the method has remarkable advantage, and can acquire higher positioning accuracy by utilizing fewer anchor nodes, and is applicable to large-scale wireless sensor networks with low anchor node density.

Description

Wireless sensor network node self positioning method based on the refinement of dividing and ruling

Technical field

The present invention is under the jurisdiction of the crossing domain that belongs to embedded development and radio communication, is specifically related to a kind of wireless sensor network node self positioning method based on the refinement of dividing and ruling.

Background technology

Wireless sensor network is a kind of emerging scientific and technical network, it links together the physical message of objective world with the transmission network, various monitoring target information in monitoring in real time, perception and the collection network distributed areas, and these information are handled, send the user who needs these information to.Wireless sensor network has characteristics such as number of nodes is big, distribution is wide, the network dynamic is strong, individual node is with low cost.Its appearance makes people obtain a kind of new way that can continue real-time monitoring of environmental, is the sensing technology of a kind of " ubiquitous ", can be widely applied to human lives's every field.Especially in military field, the self-organization of wireless sensor network and fault-tolerant ability make it can not cause the collapse of whole system because of the damage of some node, be fit to very much be applied in the abominable battlefield surroundings, comprise: monitoring enemy deployment, equipment, munitions situation, battlefield surveillance, location, battlefield loss assessment, the monitoring and the scouting of nuclear, biological attack.In addition, wireless sensor network can also be widely used between environmental monitoring and forecast, health care, Smart Home, building condition monitoring, complicated machinery monitoring, urban transportation, space exploration, large car and storehouse management, and the fields such as safety detection of airport, large-scale industrial district.

Sensor network has with legacy network visibly different purpose of design and specification requirement is arranged.Sensor network is to be purpose with perception, collection and transmission data, and intermediate node only is responsible for the forwarding of packet.Distance between adjacent node is shorter, and the multi-hop communication pattern of low-power consumption is saved power consumption, has increased the disguise of communication simultaneously, has also avoided the radio communication of long distance to be subject to the influence that outside noise disturbs.The requirement of these uniquenesses and restraining factors are that Research on sensor networks has proposed new technical problem, and network self poisoning problem is exactly one of them.Because sensor node is often taked the deployment way that dispenses at random, except that the minority sensor node, most of node in the network can not be predicted its position, and the application background of sensing network is often closely related with the node geographical position, and the perception data that does not have positional information is nonsensical.Therefore, in the application of wireless sensor network, the orientation problem of node self has also just become one of crucial problem.

At present method commonly used is to utilize the node of a small amount of known location in the sensor network to obtain the positional information of the node of other unknown positions under the technical conditions.The node of known location is called anchor node; The node of unknown position is called unknown node; The maximum distance that node can be communicated by letter is called communication radius; Anchor node is set up local coordinate system according to self-position, and unknown node calculates own relative position in the local coordinate system of anchor node according to anchor node, thereby can know self-position information.

Use actual by investigating existing localization method and combining with wireless sensor network, we think, a kind of localization method of massive wireless sensor that is applicable to must reach the requirement except positioning accuracy, also need to meet the following conditions: 1. self-organization, the sensor node deployment mode often is a random distribution; 2. Distributed localization, the information of collecting the whole network in the massive wireless sensor is difficulty very; 3. communication overhead is little, the method complexity is low, and communication is the main aspect of energy consumption, and complexity concerns is to the node hardware configuration, and both directly have influence on the life cycle of network; 4. robustness, the measurement data between the transducer has error, requires localization method to have good fault-tolerance.The present invention has considered above several conditions just, is conceived to massive wireless sensor and uses, and has proposed the wireless sensor network node self positioning method based on the refinement of dividing and ruling.

Summary of the invention

In order under the condition that does not increase hardware device and Internet traffic, further to improve positioning accuracy, make targeting scheme be fit to massive wireless sensor more, the present invention proposes a kind of wireless sensor network node locating method based on the refinement of dividing and ruling, schematic diagram comprises as shown in Figure 1: the anchor node broadcast beacon packets; Unknown node receives beacon packet, and received signal intensity is converted into distance; With the anchor node opsition dependent ordering that receives; Select two anchor nodes to constitute location combination, the estimated position of calculating unknown node in order successively; Calculate the weights of each estimated position correspondence; Calculate the final position of unknown node.

In the technique scheme, the anchor node broadcast beacon packets is specially: in the locating periodically, anchor node comprises self ID and coordinate with the maximum power broadcast beacon packets.

In the technique scheme, unknown node receives beacon packet, received signal intensity is converted into distance is specially: unknown node receives and the record beacon packet, according to RSSI range finding model corresponding signal strength signal intensity is converted into distance.

In the technique scheme, the anchor node opsition dependent ordering that receives is specially: unknown node is calculated the barycenter of the anchor node that self receives, with this barycenter is initial point, calculates the angle of each anchor node with respect to barycenter respectively, arranges each anchor node by angular dimension then.

In the technique scheme, select two anchor nodes to constitute the location combination in order successively, the estimated position of calculating unknown node is specially: according to the order that sequences, select two anchor nodes to constitute the location combination successively, calculate the estimated position of unknown node, wherein, last anchor node and first anchor node constitute the location combination.If the estimated position that certain location combination calculation goes out has two, then to introduce another anchor node and reject wrong estimated position as the checking node, the checking node must be no more than the threshold value that pre-sets with the conllinear degree of two anchor nodes that constitute the location combination.If the equation group that constitutes of certain location has do not separate, just with the barycenter of all anchor nodes of receiving as approximate solution.

In the technique scheme, the weights that calculate each estimated position correspondence are specially: the weights of each estimated position correspondence are the sum reciprocal that unknown node arrives the distance of two anchor nodes that constitute this location combination.

In the technique scheme, the final position of calculating unknown node is specially: if the anchor node beacon packet that unknown node receives is not less than 3, then locate the estimated position of combination calculation and the final position that corresponding weights calculate unknown node according to each.Otherwise, directly with the barycenter of all anchor nodes of receiving as positioning result at last.

The present invention proposes a kind of distributed localization method, the anchor node that receives is divided the combination of formation location in order in twos, and then can calculate several estimated positions of unknown node, and make full use of the information of measurement data, defined the weight calculation method of each estimated position correspondence, the final position that last weighting refinement draws unknown node.The inventive method does not need to increase hardware device, does not need to increase Internet traffic and position fixing process is simple, is convenient to realize.

Below by drawings and Examples, technical scheme of the present invention is described in further detail.

Description of drawings

Fig. 1 is the wireless sensor network node self positioning method schematic diagram that the present invention is based on the refinement of dividing and ruling;

Fig. 2 is the verification method special case schematic diagram of node estimated position among the present invention;

Fig. 3 is checking node selecting method schematic diagram among the present invention;

Fig. 4 can't draw the estimated position schematic diagram of unknown node for a certain location combination among the present invention;

Signal when Fig. 5 exists for noise in the network disturbs schematic diagram;

Fig. 6 is that the positioning accuracy of the inventive method when anchor node is evenly disposed is with average anchor node number situation of change;

Fig. 7 is that the positioning accuracy of the inventive method when anchor node is disposed at random is with average anchor node number situation of change;

Fig. 8 is that the positioning accuracy of the inventive method when anchor node is evenly disposed is with the noise level situation of change;

Fig. 9 is that the positioning accuracy of the inventive method when anchor node is disposed at random is with the noise level situation of change;

Embodiment

Step 1: in the locating periodically, the anchor node broadcast beacon packets comprises self ID and coordinate.

Step 2: RSSI range finding between node.

Because sensor node has communication module, the chip of Control on Communication generally can provide the method for measure R SSI, and receives and sends messages between node in the network and just can finish the measurement of RSSI, so RSSI is the distance measuring method of a kind of low-power, low-cost.Range finding has considerable influence to radio propagation path loss to RSSI, and propagation model in the free space (free space propagation model) is:

Loss＝32.4+10×k×log ₁₀(d)+10×k×log ₁₀(f) (1)

In the formula, d is the distance of measurement point apart from information source, and k is the path attenuation factor, and f is signal frequency (MHz).Also have logarithm in addition apart from path loss model, breathe out its model, logarithm-normal distribution model etc.Consider reasons such as multipath in the actual conditions, reflection, barrier obstruct, in the actual environment signal transmission anisotropic often, path loss is compared with theoretical value also and is changed to some extent.And logarithm-normal distribution model is incorporated into ambient noise in the calculating of received energy, so more be tending towards vague generalization, is expressed as follows:

${\left[\frac{P_r\left(d\right)}{P_r\left(d_0\right)}\right]}_{\mathrm{dB}}=-10\×\mathrm{\β}\×\log \left[\frac{d}{d_0}\right]+X_{\mathrm{dB}}---\left(2\right)$

In the formula, d ₀Be reference distance, value is 1 usually; P _r(d ₀) be reference energy; X _DBBe that average is 0 gaussian variable; β is the signal attenuation factor.By formula (2) as can be known, if do not consider that ambient noise, received signal power are the functions with the distance monotone decreasing.

Step 3: anchor node ordering.

If in the communication radius of unknown node k anchor node arranged, coordinate is respectively and is (x _i, y _i), i=1 wherein, 2 ..., k then gets the barycenter U of this k anchor node _c(x _o, y _o) as interior point, set up coordinate system, according to its angle of coordinate Calculation of each anchor node, finally press angle with the anchor node arranged clockwise with respect to interior point.If exist some anchor nodes to the identical situation of angle of interior point, promptly in a little therewith anchor node conllinear of point, then with the anchor node of conllinear by its sorting once more to interior point apart from length.

Wherein, interior point calculating method is as follows:

$x_o=\frac{\underset{i=1}{\overset{k}{\mathrm{\Σ}}}{x}_i}{k},$ $y_o=\frac{\underset{i=1}{\overset{k}{\mathrm{\Σ}}}{y}_i}{k}---\left(3\right)$

Step 4: divide and conquer is node location according to a preliminary estimate.

As shown in Figure 1, unknown node U receives A ₁～A _kBe total to k anchor node information, coordinate is respectively (x _i, y _i), i=1 wherein, 2 ..., k records U and is respectively d to the spacing of each anchor node ₁～d _k, after ordering, two anchor node A in order _iAnd A _I+1Constitute the location combination, following equation group arranged:

$\left\{\begin{array}{c}{(x_{\mathrm{ei}}-x_i)}^2+{(y_{\mathrm{ei}}-y_i)}^2=d_i^2\\ {(x_{\mathrm{ei}}-x_{i+1})}^2+{(y_{\mathrm{ei}}-y_{i+1})}^2=d_{i+1}^2\end{array}\right.---\left(4\right)$

In the formula, (x _Ei, y _Ei) estimated coordinates of unknown node when being the i time combination.Especially, A _kWith A ₁Constitute the location combination.Separate above-mentioned equation group, following three kinds of situations arranged:

(1) equation group has two to separate.At this moment, introduce except that A _iAnd A _I+1Outer another anchor node A _j, by investigating at 2 to A _jDistance recognize the estimated position of U.

If A _iAnd A _I+1Constitute the location combination, solving equation group (4) gets U _Ei(x _Ei, y _Ei) and U _Ti(x _Ti, y _Ti) 2 points.At this moment, introduce another anchor node A _jVerify, calculate U respectively _EiAnd U _TiTo A ₃Apart from d ' _EiAnd d ' _Ti:

$\left\{\begin{array}{c}{d}_{\mathrm{ei}}^{\′}=\sqrt{{(x_{\mathrm{ei}}-x_j)}^2+{(y_{\mathrm{ei}}-y_j)}^2}\\ d_{\mathrm{ti}}^{\′}=\sqrt{{(x_{\mathrm{ti}}-x_j)}^2+{(y_{\mathrm{ti}}-y_j)}^2}\end{array}\right.---\left(5\right)$

Judge that again two distances and actual range finding are from d ₃Error:

$\left\{\begin{array}{c}{\mathrm{Err}}_{\mathrm{ei}}=\left|\right|d_{\mathrm{ei}}^{\′}-d_j\left|\right|\\ {\mathrm{Err}}_{\mathrm{ti}}=\left|\right|d_{\mathrm{ti}}^{\′}-d_j\left|\right|\end{array}\right.---\left(6\right)$

Compare Err _EiAnd Err _Ti, keep smaller value, its correspondence separate the estimated position that is U.

For example, as shown in Figure 1, A ₁And A ₂Constitute the location combination, Simultaneous Equations is tried to achieve to separate and is U _E1(x _E1, y _E1) and U _T1(x _T1, y _T1) 2 points, introduce checking node A ₃, calculate U respectively _E1And U _T1To A ₃Apart from d ' _E1And d ' _T1, learn Err _T1Less than Err _E1So, judge U _T1(x _T1, y _T1) be the estimated position of U.

Can from separating, two of equation group reject false solution by above-mentioned verification method generally speaking, but owing to factors such as range finding errors between anchor node position relation and node, node and two special cases that the positioning anchor node is substantially aligned may appear verifying, as shown in Figure 2.

Among Fig. 2, A ₃And A ₄Constitute the location combination, obtain U _E3And U _T32 points are if select A ₂As the checking node, because A ₂, A ₃And A ₄Substantially in alignment on the position, cause d _E3With d _T3Differ less, if range finding occurs will drawing Err probably than mistake in addition _E3Less than Err _T3Thereby select U _E3Point is as the false judgment of estimated position.Consider this kind situation, introduce internodal conllinear degree and help select the checking node.

As shown in Figure 3, A _iAnd A _I+1Constitute the location combination, introducing conllinear degree (Degree Of Collinearity, DOC):

$\mathrm{DOC}=\cos \mathrm{\&theta;}=\frac{{\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="HSA00000269509000031.png,HSA00000269509000032.png"\; state="\{\{state\}\}">b</figure-callout>}}^2+c^2-{\mathrm{<figure-callout\; id="2"\; label="a"\; filenames="HSA00000269509000031.png,HSA00000269509000032.png"\; state="\{\{state\}\}">a</figure-callout>}}^2}{2\&times;b\&times;c}---\left(7\right)$

If A _j, A _iAnd A _I+13 conllinear degree is less than the threshold value DOC that pre-sets _m, promptly θ is not less than a certain threshold value θ _m, then can select A _jAs the checking node.DOC _mValue can suitably be adjusted according to node density in the network and noise level.

(2) equation group has only one to separate.Promptly with A _iBe center of circle d _iFor the circle of radius with A _I+1Be center of circle d _I+1For the circle of radius tangent, at this moment, directly with this point as the estimated position.

(3) equation group does not have and separates.As shown in Figure 4, A ₄And A ₅Constitute the location combination, because 2 original apart from each others, and there is error in range finding, causes the no intersection points of two circles.This kind situation can be with the barycenter U of all anchor nodes _c(x _o, y _o) as approximate solution.

In sum, by k anchor node opsition dependent being divided in proper order combination in twos, each combination can draw an estimated position U of unknown node _Ti(x _Ti, y _Ti), i=1,2 ..., k.At this moment, for unknown node U, the locating information of its storage can be by following matrix notation:

$\mathrm{Data}=\left[\begin{array}{cccc}1& (x_1,y_1)& d_1& (x_{\mathrm{<figure-callout\; id="1"\; label="t"\; filenames="HSA00000269509000021.png,HSA00000269509000023.png"\; state="\{\{state\}\}">t</figure-callout>}1},y_{t1})\\ 2& (x_2,y_2)& d_2& (x_{\mathrm{<figure-callout\; id="2"\; label="t"\; filenames="HSA00000269509000031.png,HSA00000269509000032.png"\; state="\{\{state\}\}">t</figure-callout>}2},y_{t2})\\ .& & & .\\ .& ...& ...& .\\ .& & & .\\ k& (x_k,y_k)& d_k& (x_{\mathrm{tk}},y_{\mathrm{tk}})\end{array}\right]---\left(8\right)$

Step 5: the weights that calculate each estimated position correspondence.

From the Data matrix as can be seen, unknown node calculates k estimated position by divide and conquer, and each location combination should be equal to not to the utmost to the influence of final positioning result; Further analysis, because unknown node is different to the distance of each anchor node, two anchor nodes of formation location combination should have the contribution of different sizes to the calculating of estimated position.

A _iAnd A _I+1Constitute the location combination, wherein 1≤i≤k records unknown node to A _iAnd A _I+1Distance be respectively d _iAnd d _I+1, the estimated position is U _Ti(x _Ti, y _Ti), corresponding weights are w _Ti, especially, A _kWith A ₁Constitute combination, corresponding weights are w _Tk, then

$w_{\mathrm{ti}}=\frac{1}{d_i}+\frac{1}{d_{i+1}},$ $w_{\mathrm{tk}}=\frac{1}{d_k}+\frac{1}{d_1}---\left(9\right)$

As seen, above-mentioned weight calculation method has been taken all factors into consideration euclidean distance between node pair to the influence of estimated position and the location combination influence to final positioning result, has made full use of the information of measurement data, with further raising positioning accuracy.

Step 6: the final position of calculating unknown node.

If k＜3, then directly with the barycenter of k anchor node final position as unknown node.If k 〉=3, convolution (8) and formula (9), to the final positioning result of each estimated position weighting refinement as unknown node:

$(x_e,y_e)=(\frac{\left(\underset{i=1}{\overset{k}{\mathrm{\&Sigma;}}}{w}_{\mathrm{ti}}{x}_{\mathrm{ti}}\right)}{\underset{i=1}{\overset{k}{\mathrm{\&Sigma;}}}{w}_{\mathrm{ti}}},\frac{\left(\underset{i=1}{\overset{k}{\mathrm{\&Sigma;}}}{w}_{\mathrm{ti}}{y}_{\mathrm{ti}}\right)}{\underset{i=1}{\overset{k}{\mathrm{\&Sigma;}}}{w}_{\mathrm{ti}}})---\left(10\right)$

Because the Data matrix upgrades with the variation of locating periodically, part just can be found in time because of reasons such as the depleted of energy anchor node of no longer working that breaks down, so the inventive method is healthy and strong, the minority anchor node breaks down can not have influence on the location of whole network.

Be the correctness and the validity of method among checking the present invention, following tectonic network environment carries out emulation experiment, and under equal conditions the performance to the inventive method, centroid method and weighted mass center method compares analysis.

(1) simulation parameter setting

Experiment is based on following experiment parameter and condition (more directly perceived and be without loss of generality for making simulation result, distance parameter is a unit with rice (m) all):

(1) zone, standard rectangular level land of experimental situation: 100m * 100m, the maximum communication distance of all nodes is R;

Conllinear degree threshold value DOC when (2) the checking node is selected _mBe set to π/18, i.e. θ _m=10 °.

(3) (Average Anchor Number AAN), represents the mean value of the anchor node number that each sensor node can receive in the network to average anchor node number, regulates AAN by changing R, ignores the position error of anchor node self;

(4) (Degree ofNoise DON), is defined as average and is the standard deviation sigma of 0 Gaussian Profile, the noise level in the expression monitoring of environmental to noise level.Fig. 5 (a) and (b) have reflected that respectively DON is 0.1 and the transmission situation of 0.2 o'clock signal.

(5) (Average Localization Accuracy ALA) is to define average positioning accuracy

$\mathrm{ALA}=\frac{1}{N_l\&times;R}\underset{i=1}{\overset{N_l}{\mathrm{\&Sigma;}}}\sqrt{{(x_{\mathrm{ei}}-x_{\mathrm{ai}})}^2+{(y_{\mathrm{ei}}-y_{\mathrm{ai}})}^2}---\left(11\right)$

In the formula, N _lBe oriented interstitial content, (x _Ai, y _Ai) be the physical location of unknown node, (x _Ei, y _Ei) be the estimated position of unknown node.

From positioning principle as can be known, the major parameter that influences the positioning performance of centroid method, weighted mass center method and the inventive method is average anchor node number AAN and noise level DON, below under the network topology scene that anchor node is evenly disposed and disposed at random each side's legal position Effect on Performance is carried out emulation relatively with regard to both respectively.

(2) position error is with anchor node variable density situation

Under the network environment of DON=0.1, Fig. 6 and Fig. 7 have shown that respectively the positioning accuracy of each method under the even condition of disposing and disposing at random of anchor node is with average anchor node number situation of change.As can be seen, the positioning accuracy of centroid method and weighted mass center method rises gradually with the increase of AAN, this is because the anchor node information that sensor node receives in communication range is many more, the information that can be used for locating is also just many more, and the main thought of these two kinds of methods is averaged exactly, so reduced position error from probability.And the positioning accuracy of the inventive method is for the increase and decrease of AAN obvious variation too not, analyze from its positioning principle, as long as the anchor node that can be used for locating is not less than 3, and the conllinear degree of 3 anchor nodes satisfies the check post and selects requirement, just can more accurately calculate the estimated position of unknown node, so the variation of anchor node number does not have too much influence for the inventive method.As can be seen from Figures 6 and 7, along with AAN increases gradually, the position error of the inventive method slowly descends near 10% basically, and this is because the increase of anchor node information not only makes method have more locating information, and effectively method of weighting has also played certain refinement effect.Especially, when AAN=4, for the even situation of disposing of anchor node, the positioning accuracy of the inventive method has improved 14% and 11% respectively than centroid method and weighted mass center method, and the situation of disposing at random for anchor node, the position error of the inventive method has reduced by 22% and 17% respectively than centroid method and weighted mass center method, and these group data have fully shown the inventive method and have remarkable advantages in the network positions of low anchor node density.

(3) position error is with the range error situation of change

By changing node maximum communication radius R keeping AAN=6, Fig. 8 and Fig. 9 be respectively anchor node evenly dispose and at random under the deployment scenario each side's method positioning accuracy with the situation of change of noise level DON.As can be seen, the positioning accuracy of centroid method is less with the variation of DON, maintains same level substantially, and this is because centroid method directly utilizes the internodal connective location of realizing, do not use received signal intensity, so the interference of signal is little to the legal position of barycenter Effect on Performance.And the variation of DON has produced considerable influence to the positioning accuracy of weighted mass center method and the inventive method, illustrates that signal disturbs the error that range finding is produced to come negative effect necessarily to positioning belt.See on the whole, when anchor node is evenly disposed, the position error of the inventive method has on average reduced about 10% and 2% respectively than centroid method and weighted mass center method, when anchor node was disposed at random, relative centroid method of the position error of the inventive method and weighted mass center method had on average reduced about 12% and 6% respectively.On the other hand, if DON is 0.3 with interior variation, the positioning accuracy of the inventive method still is gratifying, can remain in 18%, can satisfy the positioning requirements that great majority are used.

From positioning principle and simulation result as can be seen, no matter anchor node is evenly disposed is still disposed at random, the inventive method can both obtain comparatively accurate localization effect, and especially under the situation of low anchor node density, the positioning accuracy of the inventive method increases significantly.Secondly, the inventive method does not increase any hardware device, and the traffic is extremely low, if having M anchor node in the network, position fixing process only need transmit M packet.So the inventive method principle is simple, expense is little in the position fixing process, and the precision height is suitable for the massive wireless sensor of low anchor node density.

Claims (7)

1. based on the wireless sensor network node self positioning method of the refinement of dividing and ruling, it is characterized in that may further comprise the steps:

Step 1: in the locating periodically, the anchor node broadcast beacon packets;

Step 2: after unknown node receives the beacon packet of anchor node, received signal intensity is converted into the range information of anchor node to self;

Step 3: after unknown node received beacon packet, opsition dependent sorted anchor node;

Step 4: select two anchor nodes to constitute location combination, the estimated position of calculating unknown node in order successively;

Step 5: the weights that calculate each estimated position correspondence;

Step 6: the final position of calculating unknown node.

2. according to right 1 described wireless sensor network node self positioning method based on the refinement of dividing and ruling, it is characterized in that described step 1 is specially: in the locating periodically, anchor node comprises self ID and coordinate with the maximum power broadcast beacon packets.

3. according to right 1 described wireless sensor network node self positioning method based on the refinement of dividing and ruling, it is characterized in that, described step 2 is specially: unknown node receives and the record beacon packet, according to RSSI range finding model corresponding signal strength signal intensity is converted into distance.

4. according to right 1 described wireless sensor network node self positioning method based on the refinement of dividing and ruling, it is characterized in that, described step 3 is specially: unknown node is calculated the barycenter of the anchor node that self receives, with this barycenter is initial point, calculate the angle of each anchor node respectively, arrange each anchor node by angular dimension then with respect to barycenter.

5. according to right 1 described wireless sensor network node self positioning method based on the refinement of dividing and ruling, it is characterized in that, described step 4 is specially: 3 orders that sequence set by step, select two anchor nodes to constitute the location combination successively, calculate the estimated position of unknown node, wherein, last anchor node and first anchor node constitute the location combination.If the estimated position that the location combination calculation goes out has two, then to introduce another anchor node and reject wrong estimated position as the checking node, the checking node must be no more than the threshold value that pre-sets with the conllinear degree of two anchor nodes that constitute the location combination.If the equation group that constitutes of certain location has do not separate, just with the barycenter that calculates in the step 3 as approximate solution.

6. according to right 1 described wireless sensor network node self positioning method based on the refinement of dividing and ruling, it is characterized in that described step 5 is specially: the weights of each estimated position correspondence are the sum reciprocal that unknown node arrives the distance of two anchor nodes that constitute this location combination.

7. according to right 1 described wireless sensor network node self positioning method based on the refinement of dividing and ruling, it is characterized in that, described step 6 is specially: if the anchor node beacon packet that receives of unknown node is not less than 3, then the corresponding weight value that obtains of estimated position that obtains according to step 4 and step 5 calculates the final position of unknown node.Otherwise, directly with the barycenter of all anchor nodes of receiving as positioning result at last.

Images