CN103152791A - Target tracking method based on underwater wireless sensor network - Google Patents

Abstract

The invention discloses a target tracking method based on an underwater wireless sensor network. The method selects cluster head nodes according to a strongest signal principle at first. Then, a cluster network is formed according to single hop distance principle to observe a target. If observation signal strength exceeds a threshold value, observation data are sent to cluster head nodes. The cluster head nodes receive the data which are transmitted by thick internal nodes. A resampling particle filtering algorithm is utilized to estimate a target position and variance at present moment. The cluster head nodes are constantly updated according to target movements. A node state estimating value and a variance value of a previous cluster head node are transmitted to a present cluster head node. The present cluster head node estimates moving target positions through a modified resampling particle filtering algorithm until the moving target exceeds a tracking range of the underwater wireless sensor network. The target tracking method based on the underwater wireless sensor network can estimate the position and the variance of the underwater target by utilizing the filtering tracking method of the modified resampling particle algorithm. Target tracking property of the underwater wireless sensor network is improved.

Description

A kind of method for tracking target based on underwater wireless sensor network

Technical field

The present invention relates to a kind of method for tracking target based on underwater wireless sensor network.

Background technology

The network that underwater wireless sensor network refers to dispose a large amount of sensor nodes and Autonomous Vehicles cooperation monitoring in certain waters and gathers the surrounding environment data of interest, sensor node can be set up to self-organizing network and carry out sound communication, through Data fusion technique, specified node is sent to the data of obtaining the control centre of the water surface or bank base, has so just realized the fusion of underwater sensor network and terrestrial communications network.Underwater sensor has low-power consumption, the short characteristics of transmission range usually.

It is an important application of underwater sensor network that submarine target is followed the tracks of.Underwater sensor network has node wide, the characteristics such as quantity is many, can mutually cooperate between node, swap data, and extensibility is strong that distribute, this be conducive to enlarge target following range, strengthen reliability and the real-time of target following.

Submarine target is followed the tracks of and mostly is nonlinear problem, particle filter method is widely applied in non-linear non-Gauss's problem, existing particle filter is used for Target Tracking Problem in the wireless sensor network of land, so the submarine target tracking problem adopts particle filter method more.

The method for tracking target of underwater sensor network can be divided into centralized particle filter tracking and distributed particle filter tracking according to the difference of particle filter working method.Only have a Centroid in centralized particle filter tracking method network, all the other nodes send the measurement data of tracking target to Centroid, and Centroid is responsible for using particle filter to carry out data and is processed, and estimates the movement locus of maneuvering target.Centralized particle filter tracking method makes the stable not and laod unbalance of whole network, and distributed particle filter tracking method has overcome these shortcomings, is divided into haply four kinds according to the difference of particle filter existence form.Different pieces of information processing node operation particle filter algorithm in the same time not in the first method network according to the prediction locus of maneuvering target, is selected from the nearest feasible node of predicted position as processing node.Processing node changes along with the change of maneuvering target track, has solved the problem of the fragile and laod unbalance of network in centralized particle filter.Second method is in order to overcome the huge loss that may exist in query script, adopt the simple mode that expands processing node, be that in network, each processing node moves identical particle filter algorithm simultaneously, the data that measured value gathers from whole network sensor.The third method supposes that each processing node has the transducer of respective numbers unique connected with it, and each processing node moves different particle filter algorithms simultaneously, the data that the transducer that the employing of renewal measured value is attached thereto gathers.The 4th kind of method considers that the particle filter population is more, and be higher to the precision of maneuvering target tracking track, and single processing node finite capacity moves particle filter respectively so the particle collection is divided into several subsets in different processing nodes.

Summary of the invention

The objective of the invention is to overcome the deficiencies in the prior art, a kind of method for tracking target based on underwater wireless sensor network is provided.

Step based on the method for tracking target of underwater wireless sensor network is as follows:

1) initialization underwater wireless sensor network makes all the sensors node all have same specification, and all in running order;

2) sensor node of selecting received signal strength maximum under water is as leader cluster node, the sensor node with leader cluster node in the single-hop communication scope and leader cluster node group bunch, and all the other sensor nodes remain on resting state;

3) a bunch inner sensor node is observed target, and the intensity of the signal that receives is compared with default thresholding, if send data to leader cluster node higher than default thresholding, otherwise does not send;

4) set Initial state estimation value and initial variance estimated value;

5) at k constantly according to step 2) group bunch, and send upper one state estimation value and the estimate of variance packing estimated of particle filter constantly to this k leader cluster node constantly;

6) carry out the particle filter of k improvement resampling constantly, N particle of sampling from the importance density function, then upgrade the sampling particle, carry out particle and improve resampling, location estimation value and the estimate of variance of last export target;

7) k from adding 1, constantly upgrades leader cluster node according to the motion of target constantly, and the information with a upper leader cluster node when leader cluster node is changed sends current leader cluster node to;

8) repeating step 5)-step 7), until till the overlay area of target disengaging underwater wireless sensor network.

Described step 6) is: from

Gather particle I=1 ..., N, and calculate weights of importance and normalization weights of importance, wherein

Be that suggestion distributes, the N of collection particle weights of importance is

After normalization, weight is

To the particle renewal that resamples, update method is: select the particle that keeps according to the size of particle weight, particle passes the method that the probability wall namely is saved and is rejected originally, when the particle weight is passed a plurality of probability wall, abandon copying the pattern of particle, and adopt following strategy:

When copying number and be the 2n even number, produce 2 (n-1) individual new particle:

$\{x-\frac{2n-3}{4N}*\mathrm{\Σ}\},...,\{x-\frac{1}{4N}*\mathrm{\Σ}\}\left\{x\right\}\left\{x\right\}\{x+\frac{1}{4N}*\mathrm{\Σ}\},...,\{x+\frac{2n-3}{4N}*\mathrm{\Σ}\},$

When copying number and be the 2n+1 odd number, produce 2n new particle:

$\{x-\frac{2n-1}{4N}*\mathrm{\Σ}\},...,\{x-\frac{1}{4N}*\mathrm{\Σ}\left\}\right\{x\left\}\right\{x\left\}\right\{x+\frac{1}{4N}*\mathrm{\Σ}\},...,\{x+\frac{2n-1}{4N}*\mathrm{\Σ}\}$

Wherein ∑ is the decentralization variable of new particle.

The present invention uses the particle filter tracking method of improving the resampling algorithm to estimate position and the variance of submarine target, improves the performance of target tracking of underwater wireless sensor network.

Description of drawings

Fig. 1 is underwater sensor network pursuit movement object delineation.

Embodiment

Method for tracking target based on underwater wireless sensor network comprises the following steps:

1) initialization underwater wireless sensor network makes all the sensors node all have same specification, and all in running order;

2) sensor node of selecting received signal strength maximum under water is as leader cluster node, the sensor node with leader cluster node in the single-hop communication scope and leader cluster node group bunch, and all the other sensor nodes remain on resting state;

3) a bunch inner sensor node is observed target, and the intensity of the signal that receives is compared with default thresholding, if send data to leader cluster node higher than default thresholding, otherwise does not send;

4) set Initial state estimation value and initial variance estimated value;

5) at k constantly according to step 2) group bunch, and send upper one state estimation value and the estimate of variance packing estimated of particle filter constantly to this k leader cluster node constantly;

6) carry out the particle filter of k improvement resampling constantly, N particle of sampling from the importance density function, then upgrade the sampling particle, carry out particle and improve resampling, location estimation value and the estimate of variance of last export target;

7) k from adding 1, constantly upgrades leader cluster node according to the motion of target constantly, and the information with a upper leader cluster node when leader cluster node is changed sends current leader cluster node to;

8) repeating step 5)-step 7), until till the overlay area of target disengaging underwater wireless sensor network.

Described step 6) is: from

Gather particle I=1 ..., N, and calculate weights of importance and normalization weights of importance, wherein

Be that suggestion distributes, the N of collection particle weights of importance is

After normalization, weight is

To the particle renewal that resamples, update method is: select the particle that keeps according to the size of particle weight, originally particle passes the method that the probability wall namely is saved and is rejected, when the particle weight is passed a plurality of probability wall, abandon copying the pattern of particle, and adopt following strategy: when copying number and be the 2n even number, produce 2 (n-1) individual new particle:

$\{x-\frac{2n-3}{4N}*\mathrm{\Σ}\},...,\{x-\frac{1}{4N}*\mathrm{\Σ}\}\left\{x\right\}\left\{x\right\}\{x+\frac{1}{4N}*\mathrm{\Σ}\},...,\{x+\frac{2n-3}{4N}*\mathrm{\Σ}\},$

When copying number and be the 2n+1 odd number, produce 2n new particle:

$\{x-\frac{2n-1}{4N}*\mathrm{\Σ}\},...,\{x-\frac{1}{4N}*\mathrm{\Σ}\left\}\right\{x\left\}\right\{x\left\}\right\{x+\frac{1}{4N}*\mathrm{\Σ}\},...,\{x+\frac{2n-1}{4N}*\mathrm{\Σ}\}$

Wherein ∑ is the decentralization variable of new particle.

Embodiment

Step 101: the initialization underwater wireless sensor network, sow at random under water wireless sensor network node in environment, all nodes all have unified specification, as communication distance, detection range etc., all nodes are all in running order, keep detecting function, but can the communication close function.

Step 102: sensor node detects target, wakes the node in investigative range up.These nodes in investigative range are organized bunch on principle, namely select the node M of signal receiving strength maximum as leader cluster node, and the node with leader cluster node M in the single-hop communication scope and leader cluster node form bunch, and all the other nodes continue to keep resting states.

Step 103: a bunch interior nodes is observed target, carries out to the received signal this locality and processes, and then sends data to leader cluster node.The strength model that node receives signal is:

$z\left(k\right)=\frac{S\left(k\right)}{{[x\left(k\right)-x]}^2+{y\left(k\right)-y]}^2}+\mathrm{\ϵ}\left(k\right)$

Wherein S (k) is other acoustic pressure of target source level, x (k) and y (k) are that target is at k two-dimensional coordinate constantly, x and y are the two-dimensional coordinates of sonar sensor, and ε (k) is independent white Gaussian noise, and the observation of all underwater sensor nodes is all independently.

The signal strength signal intensity that i node receives is processed in this locality, then sends data to leader cluster node according to criterion, and this criterion is: signal z (k) and default thresholding D that k is received constantly _ThresholdCompare, if value does not send any information lower than thresholding; If value is higher than thresholding, the information that sends is to leader cluster node.Therefore, node only has as z (k) higher than thresholding D _ThresholdThe time just transmit information to leader cluster node.The measurement that leader cluster node receives from i node is:

$z_i\left(k\right)=\frac{S\left(k\right)}{{[x\left(k\right)-x_i]}^2+{[y\left(k\right)-y_i]}^2}*{\mathrm{\λ}}_i+{\mathrm{\ϵ}}_i\left(k\right),$ Wherein ${\mathrm{\λ}}_i=\left\{\begin{array}{ccc}1& \mathrm{if}& z_i\left(k\right)>D_{\mathrm{threshold}}\\ 0& \mathrm{if}& z_i\left(k\right)\≤D_{\mathrm{threshold}}\end{array}\right..$

Step 104: at k=0 constantly, according to bunch principle of the group in step 102, form initial cluster, and select leader cluster node, set Initial state estimation value and initial variance estimated value.

Step 105: at k constantly, as shown in Figure 1, according to the group bunch principle group of step 102 bunch, and send upper one particle filter state estimation value and the estimate of variance packing of calculating constantly to this k leader cluster node constantly.

Step 106: carry out the particle filter algorithm that k improvement constantly resamples, state estimation is carried out in the position of target.From

Gather particle

I=1 ..., N, and calculate weights of importance and normalization weights of importance.Wherein

That suggestion distributes.N the particle weights of importance that gathers is

After normalization, weight is

To the particle renewal that resamples, update method is: select the particle that keeps according to the size of particle weight, particle passes the method that the probability wall namely is saved and is rejected originally.When the particle weight is passed a plurality of probability wall, abandon copying the pattern of particle, and adopt following strategy:

When copying number and be the 2n even number, produce 2 (n-1) individual new particle:

$\{x-\frac{2n-3}{4N}*\mathrm{\Σ}\},...,\{x-\frac{1}{4N}*\mathrm{\Σ}\}\left\{x\right\}\left\{x\right\}\{x+\frac{1}{4N}*\mathrm{\Σ}\},...,\{x+\frac{2n-3}{4N}*\mathrm{\Σ}\},$

When copying number and be the 2n+1 odd number, produce 2n new particle:

$\{x-\frac{2n-1}{4N}*\mathrm{\Σ}\},...,\{x-\frac{1}{4N}*\mathrm{\Σ}\left\}\right\{x\left\}\right\{x\left\}\right\{x+\frac{1}{4N}*\mathrm{\Σ}\},...,\{x+\frac{2n-1}{4N}*\mathrm{\Σ}\}$

Wherein ∑ is the decentralization variable of new particle.Improve the particle filter algorithm that resamples as shown in the table:

Step 107:k is constantly from adding 1.At next constantly, when Suitable For Moving-goal Problems arrives another position, again organize bunch by step 102, when the leader cluster node of selecting with upper one constantly leader cluster node when not identical, a upper leader cluster node sends information to current leader cluster node, and the information that transmits between leader cluster node is state estimation value and the estimate of variance of a upper moment target.Constantly upgrade bunch and leader cluster node according to the motion of target, when leader cluster node is changed, the information of a upper leader cluster node is sent to when the prevariety head.

Step 108: repeating step 105-107, until till target breaks away from the underwater wireless sensor network overlay area.

Claims (2)

1. method for tracking target based on underwater wireless sensor network is characterized in that its step is as follows:

1) initialization underwater wireless sensor network makes all the sensors node all have same specification, and all in running order;

2) sensor node of selecting received signal strength maximum under water is as leader cluster node, the sensor node with leader cluster node in the single-hop communication scope and leader cluster node group bunch, and all the other sensor nodes remain on resting state;

3) a bunch inner sensor node is observed target, and the intensity of the signal that receives is compared with default thresholding, if send data to leader cluster node higher than default thresholding, otherwise does not send;

4) set Initial state estimation value and initial variance estimated value;

5) at k constantly according to step 2) group bunch, and send upper one state estimation value and the estimate of variance packing estimated of particle filter constantly to this k leader cluster node constantly;

6) carry out the particle filter of k improvement resampling constantly, N particle of sampling from the importance density function, then upgrade the sampling particle, carry out particle and improve resampling, location estimation value and the estimate of variance of last export target;

7) k from adding 1, constantly upgrades leader cluster node according to the motion of target constantly, and the information with a upper leader cluster node when leader cluster node is changed sends current leader cluster node to;

8) repeating step 5)-step 7), until till the overlay area of target disengaging underwater wireless sensor network.

2. a kind of method for tracking target based on underwater wireless sensor network according to claim 1, it is characterized in that: described step 6) is: from

Gather particle

I=1 ..., N, and calculate weights of importance and normalization weights of importance, wherein

Be that suggestion distributes, the N of collection particle weights of importance is

After normalization, weight is

To the particle renewal that resamples, update method is: select the particle that keeps according to the size of particle weight, particle passes the method that the probability wall namely is saved and is rejected originally, when the particle weight is passed a plurality of probability wall, abandon copying the pattern of particle, and adopt following strategy:

When copying number and be the 2n even number, produce 2 (n-1) individual new particle:

$\{x-\frac{2n-3}{4N}*\mathrm{\Σ}\},...,\{x-\frac{1}{4N}*\mathrm{\Σ}\}\left\{x\right\}\left\{x\right\}\{x+\frac{1}{4N}*\mathrm{\Σ}\},...,\{x+\frac{2n-3}{4N}*\mathrm{\Σ}\},$

When copying number and be the 2n+1 odd number, produce 2n new particle:

$\{x-\frac{2n-1}{4N}*\mathrm{\Σ}\},...,\{x-\frac{1}{4N}*\mathrm{\Σ}\left\}\right\{x\left\}\right\{x\left\}\right\{x+\frac{1}{4N}*\mathrm{\Σ}\},...,\{x+\frac{2n-1}{4N}*\mathrm{\Σ}\}$

Wherein ∑ is the decentralization variable of new particle.

Images