CN103619062A - Method for positioning unknown nodes in field environment wireless sensor network - Google Patents

Abstract

The invention provides a method for positioning unknown nodes in a field environment wireless sensor network. The method comprises the following steps: (1) the perceiving stage, wherein each sensor node determines the initial node positional relation through change of transmitting power of wireless signals, the number of neighbor nodes is judged, and a preparation is made for measuring parameters of the distances between nodes; (2) the positioning stage, wherein the average single hop distance of every two beacon nodes is calculated first through the distance parameters, and then the initial positions of the unknown nodes are calculated through a least square method according to the distance parameter of each node and the beacon nodes; (3) the correcting stage, wherein boundary node detection is carried out on all the unknown nodes first, the neighbor nodes are determined by the adoption of the characteristic that the neighbor number of boundary nodes is few, then the positions of the beacon nodes are reversely positioned through the positional information of the unknown nodes, the nodes with high positioning accuracy and the nodes with poor positioning accuracy are determined from the perspective of errors, and at last positional correction is conducted on the nodes with low accuracy through the good nodes.

Description

The localization method of unknown node in wild environment radio sensing network

Technical field

The present invention relates to wireless self-organization network systems technology field, especially the implementation method of wireless sensor node location, relates in particular to a kind of implementation method based on being deployed in unknown node location in wild environment radio sensing network.

Background technology

Along with microelectric technique, the high speed development of wireless communication technology, has perception, and the microsensor node of disposal ability and wireless communication ability starts to occur, and caused people's extensive concern.This sensor node can organize themselves into a network, by co-operating mode, after the information perceiving is processed, passes to wirelessly remote terminal, thereby realizes the detected object in network's coverage area is realized to monitoring at far-end.Wireless sensor network merges information world and real world, is widely used in environmental monitoring, medical treatment & health, Industry Control, the fields such as military monitoring, and the interactive mode in the deep change Liao Renyu world.

Sensor node localization technology is a basic fundamental in wireless sensor network.Increasingly sophisticated along with testing environment, and the variation of wireless sensor network application, in the network application scene of a lot of random placements, for accurately, fast, the localization method of sensor node becomes increasingly important reliably.

In recent years, the location technology of wireless sensor node has caused scientific research personnel's close attention, some scholars have launched the further investigation of wireless sensor network orientation problem, and in a series of IEEE meetings (as Infocom, MASS, ICDCS etc.), the academic conference of ACM is (as Sensys, Mobicom, Mobisys etc.) and on periodical some important achievements in research have been delivered.Corresponding research has also been done to the location technology of sensor network by the many well-known institution of higher learning of the U.S., and has delivered corresponding scholarly work.China also pays much attention to the research of this respect, has all started exploration and the research in this field.But these researchs, still in the starting stage, also have quite poor distance apart from actual needs.

Localization method in traditional wireless sensor network generally has two kinds: the localization method of (1) localization method based on distance and (2) and range-independence.Wherein, the localization method based on distance is measured two absolute distances between node by the method for various range findings, thereby the sensor node in network is positioned.Generally, by effective distance measuring technology, can obtain higher positioning precision, but these methods have higher hardware spending conventionally, and take complicated correcting system.In addition, the method for measurement based on distance is very easily subject to the restriction of environmental factor conventionally.For example conventional based on RSSI(wireless signal strength) distance-finding method to environmental factor, humiture particularly, the impact of barrier is large especially.

The main network connectedness of localization method with range-independence.This method has reduced hardware spending to a certain extent, so accurate positioning is lower slightly with respect to the localization method based on distance.But in the wireless sensor network of large scale deployment, mainly take or with the localization method of range-independence.In the localization method of all and range-independence, the most classical is to be published in 03 year article (" DV Based Positioning in Ad Hoc Networks ") in communication system, the localization method (DV-Hop) of this article based on an average hop distance.This work calculates the length of average each jumping by the jumping figure between beaconing nodes and the distance between beaconing nodes, then by unknown node, to the jumping figure information of each beaconing nodes, calculate distance, thereby by methods such as least square methods, calculate the positional information of actual unknown node.Although the enough less hardware spendings of the method energy calculate the position of each unknown node, due to the difference of actual deployment environment, the precision calculating also will differ widely.

The research motivation of the implementation method of the wireless sensor node location based on wild environment comes from green field thousand biography projects and (is deployed in the wireless sensor network system of forest, be devoted to the monitoring of forest ecological environment, and carry out the system of long-term large-scale sensor network).Change sensor network and mainly gather the various information in forest, comprise humiture, illumination, gas concentration lwevel etc., thus carry out the measuring and calculating of Yubi degree, the observation of small-sized climate change, the research of tropical island effect, the take precautions against natural calamities research of hidden danger and search and rescue aspect of forest.Therefore, the location of sensor node in forest seems particularly important.In forest environment, because having caused the complexity of wireless communications environment, barrier, the difference of vegetation caused traditional localization method based on distance to have larger error, and no longer applicable and complicated wild environment.For example wireless signal strength can be decayed because of barrier, or produces multipath effect, and the communication quality of transmission of wireless signals is often subject to the restriction of environment, and humiture also has indirectly impact to radio communication in addition.Therefore, the location of the wireless sensor node in forest is faced with many difficulties in the wild.

Summary of the invention

(1) technical problem that will solve

The object of the invention is, the implementation method of unknown node location in a kind of wild environment radio sensing network is provided, thereby realize the object of location unknown node, and then improve the positioning precision of whole radio sensing network.

(2) technical scheme

For solving the problems of the technologies described above, the invention provides the localization method of unknown node in a kind of wild environment radio sensing network, described radio sensing network comprise beaconing nodes and with the unknown node of beaconing nodes UNICOM, comprising:

S101, for each sensor node in described radio sensing network, by increasing gradually the mode of wireless transmitted power, sense the number of the neighbor node of described each sensor node, neighbor node based on feedback receives that the order of signal obtains the position relationship of described each sensor node and neighbor node, and the position relationship based on all described each sensor nodes and respective neighbor node in wireless sensor network obtains the distance parameter between any two sensor nodes;

S102, for any two beaconing nodes in described radio sensing network, the distance parameter based between described any two beaconing nodes estimates the actual physics distance between described any two beaconing nodes; And, for each beaconing nodes in each unknown node in described radio sensing network and any two beaconing nodes, according to the actual physics distance between the distance parameter between described each unknown node and each beaconing nodes and described any two beaconing nodes, calculate the preliminary position of described each unknown node in radio sensing network;

S103, the preliminary position of described each unknown node is corrected, obtain the final position of described each unknown node in radio sensing network.

Concrete, in step S103, the preliminary position of described each unknown node is corrected and is comprised:

Whether S1031, described each unknown node is carried out to the detection of boundary node, be boundary node according to described each unknown node of described detection judgement;

S1032, in the situation that being judged to be boundary node, determine that described each unknown node is with respect to the correction neighbor node number of respective beacon node, and again perform step 102 based on described correction neighbor node number, obtain the final position of described each unknown node;

S1033, in the situation that judging not as boundary node, based on described preliminary position, obtain location neighbours' sequence of described each unknown node and respective beacon node, and compare with former neighbours' sequence, calculate the proportion of subsequence total in two sequences in former neighbours' sequence, and the size based on proportion judges that described each unknown node has been node or bad node; The neighbours' sequence obtaining through S101 step when wherein, described former neighbours' sequence is netinit;

In the situation that judging that described each unknown node has been node, preliminary positional information by described each unknown node is carried out antidirection finding to the position of respective beacon node, obtain the position error of described respective beacon node, and judge based on described position error whether described good node is the higher node of positioning precision; By the node that precision is higher, the poor node of precision is proofreaied and correct, obtained the final position of described good node;

And in the situation that judging that described each unknown node is bad node, the final position information based on original beaconing nodes in radio sensing network and described good node, again to bad node execution step 102, obtains the final position of described bad node;

S1034, the final position obtaining based on step S1032 and step S1033, obtain the final position of described each unknown node in radio sensing network.

Concrete, in described step S1033 " the preliminary positional information by described each unknown node is carried out antidirection finding to the position of respective beacon node; obtain the position error of described respective beacon node; and judge that based on described position error whether described good node is the higher node of positioning precision, proofreaies and correct the poor node of precision by the node that precision is higher " and comprising:

Choose four good nodes, described four good nodes and respective beacon node are performed step to S102 and S1033 successively, obtain described respective beacon node with respect to the position error of four good nodes;

Judge whether described position error is less than first threshold; If so, judge that described four good nodes are the higher node of precision, and be new beaconing nodes by described four good Node configurations; If not, judge that described four good nodes are the poor node of precision;

Based on described new beaconing nodes and described respective beacon node, again to the poor node execution step 102 of precision, obtain the correction position of the poor node of precision;

Choose again other four good nodes, repeat the judgement of above-mentioned positioning precision height, until the node that the positioning precision of all good nodes is determined and all precision are poor is corrected.

Concrete, described S1031 comprises:

Calculate the border weight P of described each unknown node _i, and judge described border weight P _iwhether be less than weight threshold;

If so, judge that corresponding unknown node is as boundary node;

If not, judge that corresponding unknown node is as non-boundary node.

Preferably, described localization method, is calculated and is revised neighbours' number by following formula: VNC (i, j)=NC (j, i) * p _i,

Wherein, P _irepresented the border weight of i unknown node, the maximum of VNC (i, j) and NC (i, j) is the correction neighbours number with respect to node j as node i.

Preferably, the distance parameter between any two sensor nodes calculates by following formula:

JND(X _i,X _j)＝NC(X _i,X _j)UNC(X _j,X _i)，

Wherein, NC (X _i, X _j) be expressed as nodes X _jwith respect to X _ineighbor node number, work as nodes X _icommunication radius for just perceiving neighbor node X _jtime, the neighbor node number that can perceive in the communication radius of place.

Concrete, the actual physics distance between two beaconing nodes is estimated by following formula:

${\mathrm{JND}}_{\mathrm{unit}}=\frac{{\mathrm{\Σ}}_{i\≠j}\mathrm{Dis}\tan \mathrm{ce}\left(R_{k,}{R}_j\right)}{{\mathrm{\Σ}}_{i\≠j}\mathrm{JND}\left(R_{k,}{R}_j\right)}=\frac{\mathrm{\Σ}\sqrt{{(X_k-X_j)}^2+{(Y_k-Y_j)}^2}}{{\mathrm{\Σ}}_{i\≠j}\text{JND}\left(R_{k,}{R}_j\right)},$

Wherein, Distance (R _k, R _j) be beaconing nodes R _kand R _jbetween Euclidean distance.

Concrete, described in calculate the preliminary position of described each unknown node in radio sensing network and comprise:

Calculate each unknown node to the distance of beaconing nodes;

Utilize least square method to calculate the preliminary position of each unknown node.

Concrete, described localization method, is characterized in that, according to following formula, calculates each unknown node to the distance of beaconing nodes: D _i, _k=JND _unitjND (v _i, R _k).

(3) beneficial effect

Be different from background technology, utilize the inventive method, by induction institute deployment region environment, change to weigh and connect between node mutually apart from far and near index, node in the wireless sensor network of a large amount of deployment in forest is located, through calculating common neighbours' distance, after location and correction, reach the object of the unknown sensor node in location, and then improve the positioning precision of node in whole wireless sensor network, effectively strengthened the positioning precision of whole sensor nodes in wireless sensor network.

Accompanying drawing explanation

Fig. 1 is the flow chart of the localization method of unknown node in wild environment radio sensing network;

Fig. 2 is the flow chart of the localization method of unknown node in wild environment radio sensing network in an execution mode;

Fig. 3 is the part schematic flow sheet based on the localization method of unknown node in Fig. 2 wild environment radio sensing network;

Fig. 4 is the wireless aware figure of sensor node A in radio sensing network;

Fig. 5 is the mutual wireless aware figure of sensor node A and sensor node B;

Fig. 6 is sensor node B physical location and the graph of a relation of preliminary position behind location;

Fig. 7 .1 is embodiment mono-location result schematic diagram;

Fig. 7 .2 is relatively schematic diagram of embodiment mono-node absolute error;

Fig. 7 .3 is embodiment mono-location arithmetic result cumulative distribution function schematic diagram;

Fig. 8 .1 is embodiment bis-trees zone location result schematic diagrams;

Fig. 8 .2 is relatively schematic diagrames of embodiment bis-part of nodes absolute errors;

Fig. 8 .3 is embodiment bis-location algorithm result cumulative distribution function schematic diagrames.

Embodiment

For making object of the present invention, content and advantage clearer, below in conjunction with drawings and Examples, the specific embodiment of the present invention is described in further detail.Following examples are used for illustrating the present invention, but are not used for limiting the scope of the invention.

The object of this invention is to provide a kind of implementation method based on being deployed in the wireless sensor node location of wild environment, the method is from the actual environment feature of wireless sensor network disposition, and suppose that sensor node is evenly distributed on institute's surveyed area, and position immobilizes, by induction institute deployment region environment, change measurement and connect the far and near index of mutual distance between node, reach the object of the unknown sensor node in location, and then improve the positioning precision of node in whole wireless sensor network.

In order to achieve the above object, technical scheme provided by the invention and range-independence, the implementation method based on unknown node neighborhood and range-independence location, in the present invention, the position of assumed wireless sensor node immobilizes, sensor node is evenly distributed in sensing region, by neighborhood between sensing node, obtain weighing the far and near index (being distance parameter) of distance.

In initial radio sensing network, sensor node mainly comprises two kinds of nodes: unknown node and beaconing nodes, and wherein, beaconing nodes is the sensor node with accurate location information, unknown node is the inaccurate sensor node of positional information.Method provided by the invention mainly comprises three phases: perception stage, positioning stage and calibration phase, respectively corresponding diagram 2 step S101, S102 and S103.In the present invention, described wireless sensor network is by having information gathering, forward, and radio communication function, and the sensor node of simple disposal ability forms.Each sensor node is being carried out beyond one's work, then has been endowed the function of location.The communication range of described sensor node is to take the circle that sensor node is the center of circle, the circle that communication distance is radius.By increasing the mode of wireless signal transmission power, increase communication range, thus neighbor node situation around perception.

Refer to Fig. 1 to Fig. 6, lower mask body is introduced the specific implementation process of these three steps.

(1) perception stage: at step S101, for each sensor node in described radio sensing network, by increasing gradually the mode of wireless transmitted power, sense the number of the neighbor node of described each sensor node, neighbor node based on feedback receives that the order of signal obtains the position relationship of described each sensor node and neighbor node, and the position relationship based on all described each sensor nodes and respective neighbor node in wireless sensor network obtains the distance parameter between any two sensor nodes.That is to say, in perception stage, each sensor node senses the number of neighbor node around by increasing gradually the mode of wireless transmitted power, and receives that by neighbor node the order of signal obtains absolute far and near relation.The far and near relation that this mode obtains is more accurate with respect to the relation obtaining by wireless signal strength merely.As shown in Figure 4, along with the increase of sensor node A wireless transmitted power, this scheme can obtain sequencing-GCEFBD that neighbor node occurs.Ideally, if a sensor node can perceive neighbor node far away under larger transmitting power, its communication zone one covers nearer node surely.

In perception stage, can obtain a conclusion, when institute's surveyed area environment facies are in good situation, for example spacious, humiture is moderate, does not have under environment that trees or rock block, and sensor node can perceive more neighbor node.Therefore, the radio communication radius under this environment and single-hop transmission distance are relatively large.Yet in environment in the wild, the woods for example, due to trees, the effect of blocking of shrub etc., the communication distance of wireless signal will reduce along with increasing of obstacle, thereby in equally distributed network, the neighbor node number that sensor node can sense is also relatively little.Therefore, in a wireless sensor network of evenly disposing, neighbor node quantity number reflected to a certain extent the size of radio communication radius, also reflected the length of single-hop distance.

There are some to only have the interesting localization method of a small amount of UNICOM information being suggested in recent years, consider the limited communication distance of sensor node, generally can only learn the nigh information of which node.Unfortunately, simply by virtue of wireless signal strength, can not accurately judge concrete far and near relation and the actual place of the neighbours direction of neighbor node.Therefore the relation of the applicable approximate distance of the technical program and network node connectedness is estimated unknown node position in the wild.

In order to realize the location of unknown node, the technical program proposes the index of a kind of common neighbours' distances (JND) and estimates the distance between every pair of node, proposed to using common neighbours apart from JND as the distance parameter between every two sensor nodes.Common neighbours are apart from calculating a number of jumping the neighbours' sensor node perceiving in transmission range, and field, density of woods particularly, and barrier situation and humiture are taken into account.This scheme is carried out description node spacing situation by common neighbours' distance, by equation JND (X _i, X _j)=NC (X _i, X _j) UNC (X _j, X _i) can obtain.NC (X wherein _i, X _j) be expressed as nodes X _jwith respect to X _ineighbor node number, work as nodes X _icommunication radius for just perceiving neighbor node X _jtime, the neighbor node number that can perceive in the communication radius of place.Therefore, common neighbours' distance definition of a pair of node within the scope of a jumping is nodes X _jwith respect to X _ineighbor node number upper nodes X _iwith respect to X _jneighbor node number.Referring to Fig. 5, the communication radius of sensor node A just can perceive neighbor node B, and communication radius is D _aB(D _aBeuclidean distance for node A and B), time, the communication range of that node A can cover F too, G and H.Similarly, the communication radius of sensor node B also just can perceive in the situation of neighbor node A, and communication radius is D _bA(D _bAeuclidean distance for node A and B) time, the communication range of Node B also can cover D, E, G and H.Therefore, NC (A, B)=4, overlay node { B, F, G, H}, NC (B, A)=5, overlay node { A, H, G, E, D}, JND (A, B)=NC (A, B) ∪ NC (B, A)=4+5-2=7.In this equation because node G and H appear in NC (A, B) and NC (B, A) simultaneously, therefore equation need to deduct the node number repeating, so need to deduct 2 in above formula.To introduce how based on common neighbours, apart from JND, to realize the Primary Location to unknown node below.

(2) positioning stage: for more clearly explaining the technical program, at step S102 corresponding to this stage, choose any two beaconing nodes R _kand R _j, choosing unknown node is v _i.In the technical program, i, j, k is the positive integer that is more than or equal to 1.For any two the beaconing nodes R in described radio sensing network _kand R _j, based on any two beaconing nodes R _kand R _jbetween distance parameter calculate any two beaconing nodes R _kand R _jbetween actual physics distance; And for each the unknown node v in described radio sensing network _iand R in any two beaconing nodes _kand R _jeach beaconing nodes R _kor R _j, according to described each unknown node v _iwith each beaconing nodes R _kor R _jbetween distance parameter and described any two beaconing nodes R _kand R _jbetween actual physics distance, calculate the preliminary position of described each unknown node in radio sensing network.

Referring to Fig. 4, along with the increase of sensor node A wireless transmitted power, can obtain the sequencing-GCEFBD of the neighbor node appearance of sensor node A.

Traditional DV-Hop method is to find a unknown node to the shortest path of beaconing nodes, and same, the localization method based on common neighbours' distance is also to find one from unknown node v _ito beaconing nodes R _kor R _jcommon neighbours apart from minimum path.By beaconing nodes R _kor R _jaccurate location information, the present invention can extrapolate the represented actual physics distance of each common neighbours distance, its calculation expression is:

$\left[\begin{array}{c}{\mathrm{JND}}_{\mathrm{unit}}=\frac{{\mathrm{\Σ}}_{i\≠j}\mathrm{Dis}\tan \mathrm{ce}\left(R_{k,}{R}_j\right)}{{\mathrm{\Σ}}_{i\≠j}\mathrm{JND}\left(R_{k,}{R}_j\right)}\\ =\frac{\mathrm{\Σ}\sqrt{{(X_k-X_j)}^2+{\left(Y_{k,}{Y}_j\right)}^2}}{{\mathrm{\Σ}}_{i\≠j}\mathrm{JND}\left(R_{k,}{R}_j\right)}\end{array}\right]$

Distance (R wherein _k, R _j) be beaconing nodes R _kand R _jbetween Euclidean distance.Then, the common neighbours' distance of each unknown node utilization calculates the distance of each beaconing nodes, and its calculation expression is:

D _i, _k＝JND _unit·JND(v _i,R _k)

D wherein _i,kthat represent is unknown node v _ito beaconing nodes R _kdistance.

Finally by methods such as least square methods, calculate unknown node v _ipreliminary position in radio sensing network.In the present invention, optional majority is used for estimating the preliminary position of a certain unknown node to beaconing nodes, and the two pairs of beaconing nodes of take are below example, and how specific explanations estimates unknown node v _ipreliminary position.Suppose unknown node v _icoordinate be (X ₀, Y ₀), four beaconing nodes R ₁, R ₂, R ₃, R ₄coordinate be respectively (X ₁, Y ₁), (X ₂, Y ₂), (X ₃, Y ₃) and (X ₄, Y ₄), and the distance of the unknown node to four being obtained by a location algorithm beaconing nodes is respectively L1, L2, L3 and L4.

Therefore, i beaconing nodes can be expressed as to the distance of unknown node:

$\sqrt{{(X_i-X_0)}^2}+\sqrt{{(Y_i-Y_0)}^2}.$

In theory, unknown node equals measured value to the distance of beaconing nodes, Li namely, but in fact because error causes distance and is not equal to calculated value.Therefore need to estimate unknown node coordinate by least square method, thereby make the unknown node and the distance (being assumed to be Li ') of each beaconing nodes and the error sum of squares of former distance L i minimum that calculate, namely need to make (L ₁-L ₁') ²+ (L ₂-L ₂') ²+ (L ₃-L ₃') ²+ (L ₄-L ₄') ²minimum.Concrete grammar is as follows:

The position relationship of unknown node and beaconing nodes is expressed as: (X _i-X ₀) ²+ (Y _i-Y ₀) ²=L _i ², equation launches to be expressed as: (X _i ²+ Y _i ²)-2X _ix ₀-2Y _iy ₀+ (X ₀ ²+ Y ₀ ²)=L _i ².

Unknown node v _ito four beaconing nodes R ₁, R ₂, R ₃, R ₄distance, can generate four equation group as, and equation group can be write:

Thereby try to achieve the coordinate (X of nodes of locations ₀, Y ₀), obtain unknown node v _ipreliminary position.

(3) calibration phase: at step S103, this stage is for correcting the preliminary position of described each unknown node, obtain the final position of described each unknown node in radio sensing network, further correct the position of the former unknown node Primary Location being gone out.Whole calibration phase is divided into three steps: boundary node detects, good node (node that position error is less) judgement and bad node (node that position error is larger) are revised, step S1031-S1032 in difference corresponding diagram 3, step S10330-S10335, step S10336.

Boundary node detects: according to this process, all unknown node can be divided into boundary node and non-boundary node, concrete testing process is as follows, with unknown node v _i, beaconing nodes R _kand R _jfor example.

At step S1031, to described each unknown node v _icarry out the detection of boundary node, according to described each the unknown node v of described detection judgement _iwhether be boundary node.

At step S1032, in the situation that being judged to be boundary node, determine described each unknown node v _iwith respect to respective beacon node R _kor R _jcorrection neighbor node number, and again perform step 102 based on described correction neighbor node number, according to described each unknown node v _iwith each beaconing nodes R _kor R _jbetween distance parameter and two beaconing nodes R _kor R _jbetween actual physics distance, obtain described each unknown node v _ifinal position.Calculate the process of final position and describe in the preceding article, do not repeat them here.

Concrete, described S1031 comprises:

Calculate described each unknown node v _iborder weight P _i, and judge described border weight P _iwhether be less than weight threshold;

If so, judge corresponding unknown node v _ifor boundary node;

If not, judge corresponding unknown node v _ifor non-boundary node.

In the Primary Location stage of this invention, the neighbor information of node is for weighing the index of euclidean distance between node pair distance.In most of the cases, rule of thumb can draw, the position error with the node of neighbours' number is still less relatively large.Therefore, this invention is random selects an arbitrary former unknown node as setting up a tree with node.Because the residing specific position of boundary node has caused boundary node, have less or there is no next-hop node, so this invention all leaf node under this nodes records.Same operation is carried out N time repeatedly, and the size of N depends on the size of whole wireless sensor network, and the sensor network that for example 300 nodes form can repeatable operation 100 times.In reality, N value is n/4, and wherein n is expressed as the number of sensor node in this network.Then, this invention can draw by calculating the mode of weight and the border weight of each unknown node calculate by expression formula:

$P_i=\frac{N_i}{N_{\mathrm{node}}-N_{\mathrm{landmark}}}$

P wherein _irepresented i unknown node v _iborder weight.Ni represents that it is the number of times on border that node i is identified as, N _nodeand N _landmarkrepresent respectively the number of all the sensors node and the number of beaconing nodes in network.

When node is during in network center position, because border weight is relatively little, therefore it is relatively low to be considered to the probability of boundary node, therefore this invention filters out the node on non-border by the method for setting threshold, when right is less than preset weight threshold, corresponding unknown node is boundary node.

After Preliminary screening goes out boundary node, pass through arithmetic expression:

VNC(i,j)＝NC(j,i)×p _i

Calculate node i with respect to virtual neighbours' number of node j, VNC (i wherein, j) and NC (i, j) be worth the larger correction neighbours number with respect to node j as node i between the two, and for substituting original neighbours' number, this node is reorientated.Refer to unknown node v here _iwith respect to beaconing nodes R _jvirtual neighbours' number.After this process, the error of the unknown node that type is boundary node will reduce greatly.

Through said process, the position correction of the unknown node that to have completed type be boundary node.Introducing type is below the position correction process of the unknown node of non-boundary node, has been respectively node judgement and bad node regulation, respectively step S10330-S10335 and step S10336 in corresponding diagram 3.

Good node judgement: in ecotopia, the wireless signal strength between two nodes reduces along with the increase of internodal distance.For example, but in the wild environment of actual deployment, in forest environment, wireless signal strength is set, and the impact of rock or other barrier is larger, therefore or neighbours' far and near distance order in be not one and well select.Neighbours' sweep phase in preceding method, the increase step by step of the wireless signal transmission power that the far and near relation of neighbor node can be by beaconing nodes obtains.Yet along with the increase of wireless signal transmission power, the increase of radio communication radius is nonlinear, along with increased power one-level, the appearance of neighbor node may be cluster, means that may to be greater than a neighbor node sensed.In order to obtain neighbor node sequence information more accurately, this invention is used two steps to obtain: first when an energy level of radio transmitted power increase, perceived to neighbor node be placed in a group.Equally as shown in Figure 4, when the wireless signal transmission intensity of node A is set to 4, node G perceived to and put into first group.When strength increase is to 6 time, Node B, C and E are sensed, are placed into second group.Node F and D are monitored to the 8th energy level.Therefore neighbor node can be divided into three groups ((G), (B, C, E), (D, F)).In a specific region, when neighbor node perceived arriving under some specific wireless energy signal ranks, this group node can be learnt by receiving the wireless signal strength of signal with respect to the far and near relation of beaconing nodes.Second step sorts to the node in same group according to the wireless signal strength receiving.Neighbor node sequence for arbitrary node i can be expressed as Si, therefore the neighbor node sequence of node A can be expressed as (A, G, C, E, B, F, D).

Concrete, at step S10330, based on each unknown node v _ipreliminary position obtain described each unknown node v _iwith respective beacon node R _kor R _jlocation neighbours' sequence, and compare with former neighbours' sequence, calculate the proportion of subsequence total in two sequences in former neighbours' sequence, and the size based on proportion judges that described each unknown node has been node or bad node.In example below, node A is expressed as or B is expressed as a unknown node.In the localization method of common neighbours' distance of this invention, as shown in Figure 5, Node B is the down hop of node A, the neighbor node sequence S of node A _a=(E, C, F, D, B), same S _b=(H, G, F, A).According to the localization method of common neighbours' distance, the neighbor node sequence S of surrounding of node A _awhat represent is the actual far and near distance relation of neighbor node, this invention S _a' represent to be tentatively decided to be the neighbor node sequence of posterior nodal point A.

In theory, if positioning result is accurate, neighbor node sequence S _aand S _{a '}equate, yet if both occur deviation, occur position error.The same Fig. 6 that uses is as example herein, and Node B is physical location, the position of B ' for obtaining by location algorithm.Behind location the neighbor node sequence of the node A that obtains for (E, C, F, B ', D), than the initial neighbours' sequence of node A, learns that variation has occurred the sequence of node D and B.Neighbor node sequence S _band S _{b '}difference can be used as a kind of measurement index of position error.Use in the present invention η _anode and bad node have been screened.Generally, η _athe node that numerical value is large has been that the probability of node is relatively large.η _afor example, for the proportion of identical sequence in the neighbor node sequence of node i before location and behind location, S in explanation _aand S _{a '}in, identical sequence is E, C, F, so η _afor proportion, 60%.

Although good node is out selected in above-mentioned steps relatively, wherein some node exists larger position error.Therefore, these relatively good points also need further correction.In this stage, a kind of correcting method of antidirection finding of novelty is used in this invention.

At step S10331, chosen at random four nodes in node, former beaconing nodes is carried out to antidirection finding, and record the position error of former beaconing nodes.At step S10332, judge whether described position error is less than first threshold 2%.At step S10333, if so, judge that described four good nodes are the higher node of precision, and be new beaconing nodes by described four good Node configurations.At step S10334, if not, judge that described four good nodes are the poor node of precision, then choose other four good nodes, repeat this step, until when position error is less than 2%.Four nodes of now current random selection can be thought accurate point, and as new beaconing nodes.Finally, four new beaconing nodes add former beaconing nodes, to all good node execution step S102 and the operation of S1033, to realize having remained the location again of node, complete having remained a correction procedure of node.

At step S10335, after above-mentioned correction, above-mentioned, choose other four good nodes in having remained node, repeat the judgement of above-mentioned positioning precision height, until the node that the positioning precision of all good nodes is determined and all precision are poor is corrected.That is to say, after this process, all good nodes are all identified as beaconing nodes.

At step S10336, the final position information based on original beaconing nodes in radio sensing network and described good node, to bad node execution step 102, realizes the correction to described bad node again, obtains the final position of described bad node.

At step S1034, after said process, based on step S1032, obtain the final locating information that type is the unknown node of boundary node, based on step S1033, obtain the final locating information that type is the unknown node of non-boundary node, finally obtain the final locating information of all unknown node in radio sensing network, obtain the final position of all unknown node in radio sensing network.

Finally, referring to Fig. 7 .1-7.3 and Fig. 8 .1-8.3, specifically introduce the embodiment of two wireless sensor networks location that applicant carries out for the present invention, and compare with general in recent years localization method and up-to-date localization method.

First embodiment mono-carries out at Zhejiang Forestry Institute lakeside hurst, and we have disposed 50 sensor nodes at random in the woods, as shown in Fig. 7 .1.The node that four open circles in figure represent is set to beaconing nodes, is placed on respectively on four angles of network's coverage area.The result of embodiment is presented in Fig. 7 .1-7.3, and in figure, the point of black represents the physical location at node place, and square represents the estimated position of the node that the localization method of this invention use calculates.The performance of the localization method that three Fig. 7 .1-7.3 have recorded respectively this invention in woods environment, position error etc.Compare the method (Fig. 7 .2) of DV-HOP and CDL simultaneously, and calculated respectively cumulative distribution function (Fig. 7 .3).The localization method positioning precision that result shows this invention improves about 20% for up-to-date CDL method, than DV-HOP method, improve approximately 30%, and wherein the positioning result of DV-HOP is the poorest.

Embodiment bis-carries out in the forest of Zhejiang Forestry Institute.Our random placement 230 sensor nodes, as shown in Fig. 8 .1.Each node is recorded the neighbours' of surrounding of oneself far and near relation by the mode of power scan, and records common neighbours' distance of every a pair of adjacent neighbors node.Because forest is by trees, shrub, the environment of the relative complex that rock etc. form, so the transmission range of radio communication is restricted.

230 all sensor node deployments are in an approximate rectangular region, and the work of each node is mainly to gather humiture, the environmental informations such as illumination and carbon dioxide, and pass to remote terminal by the mode of multi-hop.We select four points at close four angles as beaconing nodes, and we test and compared the positioning result of up-to-date CDL.Fig. 8 .1-8.3, in same Fig. 8 .1, the point of black represents the physical location at node place, and square represents the estimated position of the node that the localization method of this invention use calculates, and circle represents the result of CDL.The performance of the localization method that three Fig. 8 .1-8.3 have recorded respectively this invention in woods environment, position error etc.Compare the method (Fig. 8 .2) of DV-HOP and CDL simultaneously, and calculated respectively cumulative distribution function (Fig. 8 .3).The localization method positioning precision that result shows this invention improves about 20% for up-to-date CDL method, than DV-HOP method, improve approximately 40%, and wherein the positioning result of DV-HOP is the poorest.

As seen from the above, be different from localization method of the prior art, utilize the inventive method, by induction institute deployment region environment, change to weigh and connect between node mutually apart from far and near index, node in the wireless sensor network of a large amount of deployment in forest is located, through calculating common neighbours' distance, after location and correction, reach the object of the unknown sensor node in location, and then improve the positioning precision of node in whole wireless sensor network, effectively strengthened the positioning precision of whole sensor nodes in wireless sensor network.

The foregoing is only embodiments of the invention; not thereby limit the scope of the claims of the present invention; every equivalent structure or conversion of equivalent flow process that utilizes specification of the present invention and accompanying drawing content to do; or be directly or indirectly used in other relevant technical fields, be all in like manner included in scope of patent protection of the present invention.

Claims (9)

1. a localization method for unknown node in wild environment radio sensing network, described radio sensing network comprise beaconing nodes and with the unknown node of beaconing nodes UNICOM, it is characterized in that, comprising:

S101, for each sensor node in described radio sensing network, by increasing gradually the mode of wireless transmitted power, sense the number of the neighbor node of described each sensor node, neighbor node based on feedback receives that the order of signal obtains the position relationship of described each sensor node and neighbor node, and the position relationship based on all described each sensor nodes and respective neighbor node in wireless sensor network obtains the distance parameter between any two sensor nodes;

S102, for any two beaconing nodes in described radio sensing network, the distance parameter based between described any two beaconing nodes estimates the actual physics distance between described any two beaconing nodes;

And, for each beaconing nodes in each unknown node in described radio sensing network and any two beaconing nodes, according to the actual physics distance between the distance parameter between described each unknown node and each beaconing nodes and described any two beaconing nodes, calculate the preliminary position of described each unknown node in radio sensing network;

S103, the preliminary position of described each unknown node is corrected, obtain the final position of described each unknown node in radio sensing network.

2. localization method according to claim 1, is characterized in that, in step S103, the preliminary position of described each unknown node is corrected and is comprised:

Whether S1031, described each unknown node is carried out to the detection of boundary node, be boundary node according to described each unknown node of described detection judgement;

S1032, in the situation that being judged to be boundary node, determine that described each unknown node is with respect to the correction neighbor node number of respective beacon node, and again perform step 102 based on described correction neighbor node number, obtain the final position of described each unknown node;

S1033, in the situation that judging not as boundary node, based on described preliminary position, obtain location neighbours' sequence of described each unknown node and respective beacon node, and compare with former neighbours' sequence, calculate the proportion of subsequence total in two sequences in former neighbours' sequence, and the size based on proportion judges that described each unknown node has been node or bad node; The neighbours' sequence obtaining through S101 step when wherein, described former neighbours' sequence is netinit;

In the situation that judging that described each unknown node has been node, preliminary positional information by described each unknown node is carried out antidirection finding to the position of respective beacon node, obtain the position error of described respective beacon node, and judge based on described position error whether described good node is the higher node of positioning precision; By the node that precision is higher, the poor node of precision is proofreaied and correct, obtained the final position of described good node;

And in the situation that judging that described each unknown node is bad node, the final position information based on original beaconing nodes in radio sensing network and described good node, again to bad node execution step 102, obtains the final position of described bad node;

S1034, the final position obtaining based on step S1032 and step S1033, obtain the final position of described each unknown node in radio sensing network.

3. localization method according to claim 2, it is characterized in that, in described step S1033 " the preliminary positional information by described each unknown node is carried out antidirection finding to the position of respective beacon node; obtain the position error of described respective beacon node; and judge that based on described position error whether described good node is the higher node of positioning precision, proofreaies and correct the poor node of precision by the node that precision is higher " and comprising:

Choose four good nodes, described four good nodes and respective beacon node are performed step to S102 and S1033 successively, obtain described respective beacon node with respect to the position error of four good nodes;

Judge whether described position error is less than first threshold; If so, judge that described four good nodes are the higher node of precision, and be new beaconing nodes by described four good Node configurations; If not, judge that described four good nodes are the poor node of precision, then choose other four good nodes, repeat this deterministic process, until there are four good nodes to be judged as the node that precision is higher;

Based on described new beaconing nodes and described respective beacon node, again to the poor node execution step 102 of precision, obtain the correction position of the poor node of precision;

Choose again other four good nodes, repeat the judgement of above-mentioned positioning precision height, until the node that the positioning precision of all good nodes is determined and all precision are poor is corrected.

4. localization method according to claim 2, is characterized in that, described S1031 comprises:

Calculate the border weight P of described each unknown node _i, and judge described border weight P _iwhether be less than weight threshold;

If so, judge that corresponding unknown node is as boundary node;

If not, judge that corresponding unknown node is as non-boundary node.

5. localization method according to claim 4, is characterized in that, is calculated and is revised neighbours' number: VNC (i, j)=NC (j, i) * p by following formula _i,

Wherein, P _irepresented the border weight of i unknown node, VNC (i, j) and NC (i, j) are worth the larger correction neighbours number with respect to node j as node i between the two.

6. localization method according to claim 1, is characterized in that, the distance parameter between any two sensor nodes calculates by following formula:

JND(X _i,X _j)＝NC(X _i,X _j)UNC(X _j,X _i)，

Wherein, NC (X _i, X _j) be expressed as nodes X _jwith respect to X _ineighbor node number, work as nodes X _icommunication radius for just perceiving neighbor node X _jtime, the neighbor node number that can perceive in the communication radius of place.

7. localization method according to claim 6, is characterized in that, the actual physics distance between two beaconing nodes is estimated by following formula:

${\mathrm{JND}}_{\mathrm{unit}}=\frac{{\mathrm{\Σ}}_{i\≠j}\mathrm{Dis}\tan \mathrm{ce}\left(R_{k,}{R}_j\right)}{{\mathrm{\Σ}}_{i\≠j}\mathrm{JND}\left(R_{k,}{R}_j\right)}=\frac{\mathrm{\Σ}\sqrt{{(X_k-X_j)}^2+{(Y_k-Y_j)}^2}}{{\mathrm{\Σ}}_{i\≠j}\text{JND}\left(R_{k,}{R}_j\right)},$

Wherein, Distance (R _k, R _j) be beaconing nodes R _kand R _jbetween Euclidean distance.

8. localization method according to claim 7, is characterized in that, described in calculate the preliminary position of described each unknown node in radio sensing network and comprise:

Calculate each unknown node to the distance of beaconing nodes;

Utilize least square method to calculate the preliminary position of each unknown node.

9. localization method according to claim 8, is characterized in that, according to following formula, calculates each unknown node to the distance of beaconing nodes: D _i,k=JND _unitjND (v _i, R _k).

Images