CN103648108A - Sensor network distributed consistency object state estimation method - Google Patents

Abstract

The invention provides a sensor network distributed consistency object state estimation method. The method, based on information transmission between observation nodes of a sensor network, enables dynamic function division to be carried out on sensor nodes in the network, and observation node sets participating in consistency state estimation are adaptively optimized and selected in real time; on the basis of a distributed maximum a posteriori theory, weighting processing is performed on object prior information and measurement information; and with the influence of covariance of state estimation errors of different observation nodes on the calculation of average consistency state taking into consideration, and effective information consistency processing is performed, the distributed state estimation accuracy of each observation node can rapidly approach to the centralized estimation accuracy, state maintenance of blind nodes on an object is guaranteed, and the cases of endless emergence of new tracks and uncertainty of the tracks and the like can be effectively prevented.

Description

Sensor network distribution type consistency Target state estimator method

Technical field

The present invention relates to the information fusion system of sensor network, relate in particular to a kind of sensor network distribution type consistency Target state estimator method, belong to sensor information process field.

Background technology

Distributed method transmits mutually and realizes resource-sharing based on effective information between node, has high fault tolerance, is easy to the advantages such as installation and expansion in catenet, therefore in the research of distributing sensor network and in applying, receives much concern.

For sensor network multi-target distributed tracking, apply, in numerous distributed state estimation methods, algorithm for estimating based on congruity theory adopts the mode of iteration, utilize the effective information of neighbor node to constantly update local estimation, thereby make each node alone in computing network the overall situation of all information after collecting estimate, such as the state estimation global mean value of all nodes.Key is, coherence method does not need TOCOM total communication just can realize the consistent state estimation of the whole network, and approximate convergence is in centralized estimated result.Therefore, based on conforming estimation framework, the communication topology of network is not had to specific (special) requirements, in principle, be applicable to the sensor network of any random connection.In addition; in complete distributed multiple target tracking application; although some node cannot cannot detect certain target to all targets or in certain period by direct detection; but consider and between node, need cooperative cooperating; each node is necessary to keep the state estimation to all targets, otherwise will constantly occur the practical problems such as chaotic, the new flight path of flight path emerges in an endless stream.Node obtains after the measurement of target, also need to realize the correctly interconnected of measurement and flight path based on the historical flight path of all targets.And estimate that in consistency, in framework, each node is keeping the state estimation of all targets, this specific character from being applicable in essence complete distributed multiple target tracking application very much.

Traditional Kalman's consistency filtering (Kalman consensus filter, KCF) etc. coherence method supposes that all nodes all can observe each target, the contribution weights that all nodes in network are estimated overall mean state are considered as identical, and ignored the impact of different node state evaluated error covariances when calculating average homogeneity sexual state, under the scene that network is not exclusively communicated with or blind node is more, can badly influence estimated accuracy and the convergence rate of algorithm.In addition, in modern war and actual life, the sensing ability of microsensor is distance limited and anisotropic (such as video sensor, directional microphone, radar etc.) normally, and the performance of transducer relies on distance and the direction between observation station and target simultaneously.Therefore, the node in network can not keep the observation to all targets at any time, in network, exist at any time cannot the detection of a target blind node.And the change of sensing model will have influence on the target observation quality of node, to network detection covering, node Collaborative Control, Target state estimator distributed is followed the tracks of key technology and is brought new challenge.

In the target following practical application of large-scale sensor network, the single moment only has minority node can observe the target of passing through monitored area, and node in network all has the routing channel of upwards reporting monitoring information conventionally, therefore, in the limited catenet of distance, the single moment only needs minority observer nodes to realize the accurate flight path that can know target to the distributed state estimation of target, without all node TOCOM total communications, just can meet the practical application of target following.

In fact, along with target constantly moves in monitored area, observer nodes is dynamic change, by carrying out active data transmission between node, just can obtain the information about firms of single moment observer nodes.In addition, for the limited complete distributed consensus Target state estimator of the sensor network problem of distance, can estimate (maximum a posteriori based on distributed maximum a posteriori, MAP) theory, the information view mass metering of combined sensor and state estimation quality, through enough consensus information transmission and iterative processings repeatedly, the distributed precision of state estimation of each observer nodes can approach centralized optimum kalman filter method (centralized Kalman filter, CKF).This is thinking of the present invention source namely.

Summary of the invention

The object of the present invention is to provide a kind of high-precision dynamic self-adapting distributed consensus method for estimating state.In order to achieve the above object, the present invention is based on the information transmission between observer nodes in sensor network, proposed a kind of adaptive information weighting coherency state method of estimation based on distributed maximum a posteriori probability, schematic diagram as shown in Figure 1, comprising: sensor node obtains aim parameter measurement information; Network node Partition of role; Set up consistency set of node; Calculate local parameter of consistency; Observer nodes consensus information is processed and is merged; Target state estimator; Dbjective state prediction.

Technical scheme and concrete implementing measure:

For the ease of setting forth, do following agreement:

Communication connection in sensor network between node can be by non-directed graph

represent, wherein S={S ₁, S ₂..., S _ncomprised summits all in figure, represent the communication node in network, and gather

comprise limits all in figure, represented the feasible communications conduit that in network, different nodes are set up.In addition, with

represent all and S _ithere is the set of the node of direct communication connection,

in each node and S _ia certain limit in pie graph is all S _ineighbor node.Might as well suppose individual node S _ionly there is a transducer, at t, constantly observe target, S _ibe called observer nodes, its measurement can be expressed as m wherein _ifor transducer S _imetric data dimension, its state and measurement model can be expressed as

x _t+1＝Φx _t+w _k，k＝0，1，2，...， (1)

z _i＝H _ix _t+v _i，k＝0，1，2，...， (2)

Wherein, for state-transition matrix, process noise

for transducer S _ican time become observing matrix,

for the white Gaussian noise of zero-mean, variance is

order for measurement information matrix, equally also can time become.It is pointed out that observing matrix H _ibe not row full rank, have a m _i< p.Transducer S _ierror variance about target prior estimate is expressed as its information matrix is defined as

It is pointed out that the present invention for problem be not the state x of hypothetical target _ifor each z _ibe all observable, but consider in whole network

for dbjective state x _tthe situation with observability, network covers completely, and the single moment has at least a transducer can observe the target of monitored area.In addition, the communication radius that might as well suppose all the sensors node in network is not less than 2 times of sensing radius, this means, the node that simultaneously observes target is neighbours each other certainly, only needs the step communication can mutual transmission of information.The object of method provided by the present invention is, for sensor network target tracking problem, along with target constantly moves, network node member's role constantly changes, by effective information, transmit and process, realize observer nodes and the distributed consensus state estimation of the blind node of neighbours to target in any single filtering constantly.

The sensor node S of target T will constantly be observed with t below _ifor example, the concrete steps in technical scheme and execution mode are described in detail.

1. sensor node obtains aim parameter measurement information

Sensor node amount to obtain measurement information refers to by target echo and obtains measuring z about this locality of target _iwith measurement information matrix B _i, wherein

r _ithe variance of the zero-mean white Gaussian noise of measure obeying for transducer, subscript i is the identify label of transducer.

2. network node Partition of role

According to sensor node in network, whether observe the role that target and node are served as when the Target state estimator, the node that t is detected constantly to target is called observer nodes, the neighbor node of all observer nodes (not detecting target) is called blind node, and other nodes that do not detect target are called sleeping nodes.Wherein, observer nodes is carried out main state estimation work, and blind node keeps and upgrades local dbjective state by receiving the information of observer nodes, to prevent the occurring phenomenons such as new flight path emerges in an endless stream, flight path is not clear in flight path handshaking.

3. set up consistency set of node

T is node S constantly _idetect after target, by the information exchange between observer nodes, set up this and constantly keep the sensor node collection that dbjective state consistency is estimated, be called consistency set of node, be designated as C.Concrete implementing measure is:

(1) transducer S _iget the t measurement z of target T constantly _iand measurement information matrix B _i;

(2) transducer S _ibroadcast t-1 is the state estimation value of target T constantly

and self identification (ID), if T is fresh target, broadcast the measuring value about T;

(3) transducer S _ireception is from the broadcast message of neighbours' observer nodes;

(4) transducer S _iset up consistency set of node C, comprise t all observer nodes and the blind node of neighbours thereof constantly, wherein, the number of observer nodes is N '.

Due to the neighbours each other of all nodes in consistency set of node, by a step information, transmit, each observer nodes can receive that other observer nodes are about the state information of target.Meanwhile, in consistency set of node, the blind node of neighbours of all observer nodes also can be known target status information, and these dbjective states from different observer nodes reach unanimity, and this will set forth in follow-up explanation.

4. calculate parameter of consistency

Without loss of generality, the priori state of target T before t filtering is constantly designated as

prior information matrix is designated as

transducer S _iparameter of consistency

and

account form as follows:

$U_i^0=\frac{1}{N^{\&prime;}}{Y}_i^{-}\left(t\right)+H_i^T{B}_i{H}_i---\left(3\right)$

$u_i^0=\frac{1}{N^{\&prime;}}{Y}_i^{-}\left(t\right)x_i^{-}\left(t\right)+H_i^T{B}_i{z}_i---\left(4\right)$

5. consensus information is processed and is merged

Transducer S _icalculate parameter of consistency

and afterwards, combining information transmission, will and

carry out unification processing.Setting consistency iterations is K, and consistency rate factor is ζ, and concrete implementing measure is:

Three steps are from k=1 below, to k=K, and loop iteration K time

(1) to neighbor node

send parameter of consistency

with

(2) receive neighbor node

parameter of consistency with

(3) based on congruity theory, update consistency parameter:

Finally, through K iteration, transducer S _iobtain parameter of consistency and

limit is according to congruity theory, if the number of times of iteration is abundant, the parameter of consistency that each observer nodes obtains will reach unanimity gradually.

6. Target state estimator

Transducer S _icomplete after consensus information processing, utilize following formula to estimate dbjective state:

${\hat{x}}_i^{+}\left(t\right)={\left(U_i^K\right)}^{-1}{u}_i^K---\left(7\right)$

${\hat{Y}}_i^{+}\left(t\right)=N^{\&prime;}\&CenterDot;U_i^K---\left(8\right)$

Other sensor nodes in consistency set of node C are all realized the t estimation to dbjective state constantly according to above formula.

7. dbjective state prediction

Bonding state equation, the sensor node in consistency set of node predicts dbjective state and information matrix respectively, concrete accounting equation is as follows:

${\hat{x}}_i^{-}(t+1)=\mathrm{\&Phi;}{\hat{x}}_i^{+}\left(t\right)---\left(9\right)$

$Y_i^{-}(t+1)={(\mathrm{\&Phi;}{\left(Y_i^{+}\left(t\right)\right)}^{-1}{\mathrm{\&Phi;}}^T+Q)}^{-1}---\left(10\right)$

More than describe the information transmission of certain observer nodes and concrete mode and the step of processing in single filtering constantly, through effective consistency on messaging, processed, realized the distributed consensus state estimation of all the sensors to target in observer nodes set.And at the establishment stage of consistency set of node, each observer nodes is broadcast to the state value of previous moment target in network, has realized the blind node of observer nodes and neighbours thereof the consistency of dbjective state has been estimated.So design, is convenient to next transducer measurement constantly and has the interconnected of targetpath, and effectively preventing because target fast reserve causes following the tracks of failed situation, has avoided the phenomenons such as new flight path in network emerges in an endless stream, flight path is not clear.

Compared with prior art, the present invention has following beneficial effect:

(1) estimated accuracy is high

Compare with traditional consistency method of estimation, the present invention is based on distributed maximum a posteriori probability theoretical, utilize the information transmission between sensor node, by introducing information weighting strategy, set up contacting between transducer observation information quality and consensus information processing.Compare with traditional consistency method of estimation, the estimated accuracy of distributed consensus method provided by the present invention is closer to centralized optimal estimation method.

(2) state consistency efficiency is high

For the Target state estimator in the single moment, method proposed by the invention is only carried out consensus information processing between minority observes the sensor node of target, in conjunction with effective consistency rate factor, only need minority several times iteration can realize among a small circle in the quick uniform convergence of all the sensors to Target state estimator, also improved the real-time of whole state estimation procedure simultaneously.

(3) tracking reliability is high

Method proposed by the invention has realized each observer nodes and the consistency estimation to dbjective state of the blind node of neighbours thereof constantly, be conducive to next transducer measurement constantly and the data interconnection that has targetpath, and effectively prevent from causing following the tracks of failed situation because of target fast reserve, avoid the phenomenons such as new flight path in network emerges in an endless stream, flight path is not clear, improved target following reliability.

(4) calculating is low with communication energy consumption

In single filtering constantly, the consistency method of estimation that the present invention proposes only produces communication and calculates energy consumption between minority observes the node of target, and without TOCOM total communication and consistency iteration, so energy consumption is low, network cost is little, is more applicable for the sensor network of energy constraint.

(5) flexible design, realization are simply

The coherency state method of estimation that the present invention proposes can come self adaptation to adjust consistency set of node by simple information transmission between sensor node, without the prior information of precognition network members, has the feature of flexible design; Meanwhile, coherency state method does not relate to complicated algorithm, under existing hardware technology level, can directly by digit chip, adopt digital method to realize, and hardware performance is not had to special requirement, is convenient to integrated chip.

(6) applied widely

On the one hand, the same with traditional coherency state method of estimation, the present invention has the scope of application comparatively widely, can be used for small-sized wired fixed sensor networks such as video, radar; On the other hand, because this consistency method of estimation can be adjusted working node at each sampling instant real-time adaptive, extensibility is good, and this makes the present invention be particularly useful for having the large-scale wireless sensor network of dynamic topological structure.

Accompanying drawing explanation

Fig. 1 is that coherency state is estimated flow chart.

Fig. 2 is that sensor network target is followed the tracks of simulating scenes figure.

Fig. 3 is the network topological diagram that 50 sensor nodes form.

Fig. 4 is that the root-mean-square error mean value of each method of estimation is with the variation diagram of consistency iterations.

Embodiment

Below in conjunction with drawings and Examples, the present invention is described in further detail, monotrack example of simulation system when embodiment 1 is a network all standing, for describing detailed process and the systematic function of the sensor network coherency state method of estimation (being designated as DMAP-KCF) of the present invention's proposition in detail.

Embodiment 1

Target following simulating scenes as shown in Figure 2, the planar rectangular that monitored area is 100m * 100m, sensor node random distribution, target moves arbitrarily in region.The present embodiment is intended the distributed consensus method for estimating state DMAP-KCF of relatively the present invention's proposition and the performance of Kalman's consistency filtering method KCF, centralized optimum kalman filter method CKF, adopts 50 Monte Carlo simulations to get average and obtains testing comparative result.Below provide concrete dbjective state parameter, sensor parameters and parameter of consistency.

Dbjective state parameter

Target arbitrarily moves in monitored area, and its initial position and velocity attitude are definite at random, and velocity magnitude is random in [1m/s 3m/s] is interval to be determined.Dbjective state can be described as x=(x, y, v _x, v _y) four-dimensional vector, wherein (x, y) is target location, (v _x, v _y) be speed.The motion model of target is suc as formula shown in (1), and wherein, process noise variance is set to Q=diag (5,5,1,1).

For Target state estimator model, adopt equally the model shown in formula (1) and state-transition matrix and process noise.The initial time of target travel, is made as consistent by each observer nodes to the initial condition of target and initial variance.Wherein, initial prior estimate error variance is set to the true initial condition value of target is added to random noise that a zero-mean variance is is as initial priori state

Sensor parameters

50 sensor nodes are randomly dispersed in monitored area, all the sensors isomorphism, and communication radius is 41m, and sensing radius is 20m, and its network topology structure is as shown in Figure 3; Transducer can observe any in sensing radius with interior target, and can find range and angle measurement simultaneously, measure as bivector.Measuring value z _iby the linear measurement model shown in formula (2), determined, wherein noise variance R _i=10I ₂, I wherein ₂for two-dimentional unit vector.Observing matrix H _iand state-transition matrix Φ is defined as

$H_i=\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& 1& 0& 0\end{array}\right],\mathrm{\&Phi;}=\left[\begin{array}{cccc}1& 0& 1& 0\\ 0& 1& 0& 1\\ 0& 0& 1& 0\\ 0& 0& 0& 1\end{array}\right]$

Parameter of consistency

The processing stage of the consensus information of state estimation, if do not particularly not pointed out, the consistency iterations K=5 of acquiescence is set, consistency rate factor is set to

wherein m is the number of observer nodes in current filtering constantly.

Analysis of simulation result

Fig. 4 has provided the average estimated accuracy of DMAP-KCF, KCF and CKF with the situation of change of consistency iterations.At this, iterations K rises to 10 gradually by 1, and other parameters are default value.Wherein, average estimated accuracy is defined as the estimation root-mean-square error mean value in all moment, can be expressed as

$\mathrm{ARMSE}=\frac{1}{N_f}\underset{k=1}{\overset{N_f}{\mathrm{\&Sigma;}}}\sqrt{{\left|\right|\stackrel{\&OverBar;}{x}\left(k\right)-\mathrm{\&xi;}\left(k\right)\left|\right|}^2}$

Wherein, N _ffilter length, ξ (k) is the k actual position of target constantly,

for the state estimation average of k moment observer nodes to target, can be expressed as

$\stackrel{\&OverBar;}{x}\left(k\right)=\frac{1}{N^{\&prime;}}\underset{i=1}{\overset{N^{\&prime;}}{\mathrm{\&Sigma;}}}{\hat{x}}_i\left(k\right)$

Wherein, N ' is the k number of observer nodes constantly, for the state estimation of k moment transducer to target.

As seen from Figure 4, during K=1, DMAP-KCF has higher estimated accuracy than KCF, and close to centralized approach CKF.Along with the continuous increase of K, the root-mean-square error of DMAP-KCF reduces gradually, and Fast Convergent is to CKF estimated accuracy.K >=8 o'clock, the precision of two kinds of distributed algorithm for estimating tends to be steady.Generally, DMAP-KCF has promoted distributed estimated accuracy greatly with respect to KCF.Trace it to its cause, mainly because method DMAP-KCF provided by the invention has considered the priori redundant information between different observer nodes, distributed suitable weights to prior information and measurement information, realized consistent raw estimated accuracy and approached fast centralized optimum kalman filter method.

Claims (1)

1. a sensor network distribution type consistency Target state estimator method, a kind of Distributed filtering method for target following, by sensor node amount to obtain measurement information, network node Partition of role, set up consistency set of node, calculate local parameter of consistency, the processing of observer nodes consensus information and fusion, Target state estimator, dbjective state prediction realize in sensor network part preferably node in real time dbjective state is kept to dynamically consistent accurate estimation;

Wherein, sensor node amount to obtain measurement information refers to by target echo and obtains measuring z about this locality of target _iwith measurement information matrix B _i, wherein

r _ithe variance of the zero-mean white Gaussian noise of measure obeying for transducer, subscript i is the identify label of transducer; Network node Partition of role refers to, according to sensor node in network, whether observe the role that target and node are served as when the Target state estimator, the node that t is detected constantly to target is called observer nodes, the neighbor node of all observer nodes (not detecting target) is called blind node, other nodes that do not detect target are called sleeping nodes, wherein, observer nodes is carried out main state estimation work, and blind node is by receiving the information of observer nodes and keep and upgrading local dbjective state; Set up consistency set of node and refer to, observer nodes sends state information bag (the state estimation value that contains target previous moment

and the identify label of observer nodes), observer nodes and blind node receive the state information bag from neighbor node, the local dbjective state of blind node updates, all observer nodes of current time and all blind node consistence of composition sets of node, transducer in set of node is remaining the state estimation of target, wherein t constantly observer nodes add up to N '; Calculate local parameter of consistency and refer to target prior information and measurement information are weighted to processing, calculate local consensus information matrix

with local consensus information vector

account form is respectively $U_i^0=\frac{1}{N^{\&prime;}}{Y}_i^{-}\left(t\right)+H_i^T{B}_i{H}_i,u_i^0=\frac{1}{N^{\&prime;}}{Y}_i^{-}\left(t\right)x_i^{-}\left(t\right)+H_i^T{B}_i{z}_i,$ Wherein,

for prior estimate information matrix,

p _i(t) be state estimation error variance, H _ifor measurement matrix; Observer nodes consensus information is processed with fusion and is referred to, from k=1, start to k=K, by following three step loop iteration K time: observer nodes send the consensus information bag that contains local parameter of consistency and identify label, observer nodes reception from the consensus information bag of neighbours' observer nodes, upgrade local parameter of consistency

wherein, after each iteration is complete, carry out k=k+1, K is consistency iterations, and ζ is consistency rate factor, and final, each observer nodes obtains approximately uniform parameter of consistency with

target state estimator refers to, in conjunction with parameter of consistency and observer nodes number, dbjective state and information matrix upgraded,

wherein

the target estimated state obtaining for constantly filtering of t,

for corresponding state estimation information matrix; Dbjective state prediction refers in conjunction with current time Target state estimator and motion model, state and the information matrix thereof of next moment target predicted,

wherein, for t+1 dbjective state prediction constantly, Φ is state-transition matrix,

for the t+1 prediction of Target state estimator information matrix constantly, Q is the process noise variance in target movement model.

Images