CN105572636A - Underwater acoustic sensor network positioning method adapted to mobility - Google Patents

Abstract

The invention discloses an underwater acoustic sensor network positioning method adapted to the mobility. The positioning method belongs to the positioning algorithm field of the underwater acoustic sensor network, and is characterized in that the inherent mobility of nodes is adapted to effectively and the positioning precision is improved. According to the positioning method, a mathematical model is established to calculate positioning errors generated by different reference node sets, a reference node set with the minimal positioning error is selected as the optimal reference node set, another mathematic model is established, the reference message probability of all the different reference nodes received in a prescribed period of a positioned node is maximized to optimize reference message emission time of the different reference nodes, and the positioning precision is improved.

Description

The ambulant water sound sensor network localization method of a kind of adaptation

Technical field

The invention belongs to the location algorithm field of water sound sensor network, it is characterized in that adapting to the intrinsic movability of node preferably, improve positioning precision.In localization method of the present invention, first calculate by founding mathematical models the positioning error that different reference mode set produces, and then the optimum reference mode set that regioselective error is minimum, then founding mathematical models, utilize to maximize and in stipulated time section, received the probability of the reference message of all different reference modes to optimize the reference message transmission moment of different reference mode by location node, improve positioning precision.

Background technology

Water sound sensor network is the underwater monitoring network system be made up of the sensor node with acoustic communication and computing power, it is deployed in the underwater environments such as ocean, realizing the contamination monitoring of underwater environment, aquatic organism sample collection, disaster is prevented, and the aspects such as assisting navigation possess wide application prospect.In a lot of practical application, data need just have practical significance in conjunction with its geographical location information, and therefore location technology is important foundation and the support technology of water sound sensor network, most important to the design and devdlop of water sound sensor network.

Research and design applicable water sound sensor network own characteristic, energy efficient, simply accurate, extendible node positioning method are the emphasis of water sound sensor network research.The interpretational criteria of localization method quality generally has positioning precision (average localization error), anchor node to dispose ratio, the average Connected degree of network, network size and energy and computing cost etc., and wherein positioning precision weighs the most important interpretational criteria of localization method performance.

In the node location process of location water sound sensor network, select positioning error scope that different reference mode set produces by different.The positioning error scope supposing each length is identical, and with an ellipse representation, then actual location error range is by the overlapping region (as shown in Figure 1) of the positioning error scope of three length in location node place, as can be seen from Figure due to the difference that reference mode is selected, the error range in 101 (examples 1) is greater than 102 (examples 2).The present invention proposes how to select optimal reference mode set, to improve positioning precision.

Due to transmission range longer in water sound sensor network and propagation delay, by location node when the reference message of wait-receiving mode reference mode occur mobilely to impact positioning precision.In water sound sensor network, by location node receiving first reference mode in the reference message time gap of last reference mode, the movement occurred directly can affect positioning precision (as shown in Figure 2).Ideally, wished that the reference message that simultaneously can receive all reference modes is to reduce the impact of node mobility by location node.The present invention proposes to optimize the x time that reference mode sends reference message, to improve positioning precision.

Summary of the invention

Object of the present invention:

In water sound sensor network, positioning precision weighs the most important interpretational criteria of localization method performance, and the intrinsic movability of node can affect positioning precision, and then affect the performance of localization method.

The present invention improves positioning precision and the performance of localization method in water sound sensor network by the reference message transmission moment selected the intrinsic ambulant reference mode set of applicable node and optimize different reference mode.

Technical scheme of the present invention:

In water sound sensor network, for for location node, the selection scheme of its reference mode set has multiple, and the positioning error scope selecting different reference mode set to produce is different (as shown in Figure 1), and due to underwater sound sensor nodal pricing costliness, the interstitial content in water sound sensor network usually can not be a lot.

The present invention proposes by traveling through all possible reference mode set, and the error range of more different reference mode set, finds the localization method of optimum reference mode set, and gives a kind of mathematical modeling scheme, to reduce positioning error, improves positioning precision.

The present invention proposes to utilize to maximize and in stipulated time section, is received the probability of the reference message of all different reference modes to optimize the method in the reference message transmission moment of different reference mode by location node, and give a kind of mathematical modeling scheme, to reduce by location node in the time slot of the reference message of the reference message to last reference mode that receive first reference mode, the impact that the intrinsic movability of node produces, provides positioning precision.

Beneficial effect of the present invention:

In water sound sensor network, positioning precision weighs the most important interpretational criteria of localization method performance, and the influence factor of positioning precision has multiple, movability as intrinsic in node under water, the selection etc. of reference mode.

1) the present invention carrys out the positioning error scope of more different reference mode set by founding mathematical models, finds the optimum reference mode set that positioning error scope is minimum, effectively can reduce positioning error, improves positioning precision and the performance of localization method.

2) the present invention utilizes maximization to be obtained x time by the probability that location node receives the reference message of all different reference modes in stipulated time section with the reference message optimizing different reference mode by founding mathematical models, effectively can reduce the ambulant impact that node is intrinsic, improve positioning precision.

Accompanying drawing explanation

Fig. 1 reference mode Resource selection schematic diagram

Fig. 2 x time optimizes schematic diagram

In Fig. 1,101 is example 1, and 102 is example 2, and 103 represent the positioning error scope in example 1, and 104 represent the positioning error scope in example 2, and 105 is reference mode legend, and 106 is by location node legend.

In Fig. 2,201 is example 1, and 202 is example 2, and 203 is reference mode legend, and 204 is by location node legend.

Embodiment

Shown below is a kind of positioning error range size by more different reference mode set and determine the mathematical model scheme of optimum reference mode set and a kind of by maximizing the mathematical model scheme being optimized the x time of the reference message of different reference mode by the probability that location node receives the reference message of all different reference modes in stipulated time section, in order to effectively to reduce the intrinsic ambulant impact of water sound sensor network interior joint, improve positioning precision.

The present invention is not limited only to this mathematical model scheme; every when the localization method research carrying out water sound sensor network; utilize localization method thinking of the present invention; namely the objective function by setting up positioning error scope is selected optimum reference mode set or is utilized objective function to optimize the reference message transmission moment of different reference mode, all within protection scope of the present invention.

The embodiment of mathematical model scheme is described below:

A) founding mathematical models, calculates the positioning error range size that different reference mode set produces, and selects optimum reference mode set.

The locating measurement method of this programme, for TDOA (TimeDifferenceofArrival), calculates the CRLB (Cramer-RaoLowerBound) of the positioning error size of different reference mode set:

Supposing, is η=[x by the position vector of location node _r, y _r], the position vector of reference mode is ξ _n=[x _n, y _n], n is reference mode index.Objective function measuring amount is used represent, Δ t _n1represent that the reference message of the 1st reference mode and the reference message of the n-th reference mode arrive by the mistiming of location node, ω _nrepresent measurement noises, then can be expressed as,

Formula (1-1) can be converted into a vector matrix form, as follows:

Solve the nonlinear estimation problem of formula (1-2), interative least square method method of estimation can be adopted, suppose based on the estimated value η i-th iteration ⁱ, at the estimated value η of the i-th+1 time ⁱ⁺¹can be updated to:

Wherein, the R in formula (1-3) is the covariance matrix of ω, and J is Jacobian matrix, and it is:

$J=\frac{\∂h(\mathrm{\η}\·\mathrm{\ξ})}{\∂\mathrm{\η}}{|}_{\mathrm{\η}={\mathrm{\η}}^i}---(1-4)$

J can be extended to:

$J={\left(\begin{array}{cc}\frac{x_r-x_2}{d_2}-\frac{x_r}{d_1}& \frac{y_r-y_2}{d_2}-\frac{y_r}{d_1}\\ \frac{x_r-x_3}{d_3}-\frac{x_r}{d_1}& \frac{y_1-y_2}{d_2}-\frac{y_r}{d_1}\\ \·& \·\\ \·& \·\\ \·& \·\\ \frac{x_r-x_N}{d_N}-\frac{x_r}{d_1}& \frac{y_r-y_2}{d_2}-\frac{y_r}{d_1}\end{array}\right)}_{\mathrm{\η}={\mathrm{\η}}^i}---(1-5)$

Wherein, $d_1=\sqrt{x_r^2+y_r^2+z_r^2},d_n=\sqrt{{(x_r-x_n)}^2+{(y_r-y_n)}^2+{(z_r-z_n)}^2}.$ After the iteration of sufficient amount, square error (MSE) matrix can be tried to achieve, as shown in formula (1-6):

E[(η ⁱ-η) ^T(η ⁱ-η)]＝(J ^TR ^-1J) ^-1(1-6)

Suppose that each location needs r reference mode, and need s node to can be used as reference mode (if located space is n dimension, then needing n+1 reference mode) altogether, then have plant combination.For each combination, calculate and eachly tried to achieve average by the MSE of location node positioning error, the minimum combination of average will be selected as reference mode set.

B) founding mathematical models, is obtained x time by maximization by the reference message that the probability receiving the reference message of all different reference modes of location node in stipulated time section optimizes different reference mode.

Suppose to determine to be needed 3 reference modes by the position of location node, and from 3 reference modes to the distance of all nodes, with reference to the time needed for message transmissions be:

$\mathrm{\τ}=\left(\begin{array}{cccc}{\mathrm{\τ}}_{11}& {\mathrm{\τ}}_{12}& \·\·\·& {\mathrm{\τ}}_{1N}\\ {\mathrm{\τ}}_{21}& {\mathrm{\τ}}_{22}& \·\·\·& {\mathrm{\τ}}_{2N}\\ {\mathrm{\τ}}_{31}& {\mathrm{\τ}}_{32}& \·\·\·& {\mathrm{\τ}}_{3N}\end{array}\right)---(1-7)$

The reference message transfer time point supposing 3 reference modes is T=[t ₁, t ₂, t ₃], then the reference message of 3 reference modes is time of arrival on each node:

$\mathrm{\τ}=\left(\begin{array}{cccc}{t}_1+{\mathrm{\τ}}_{11}& t_1+{\mathrm{\τ}}_{12}& \·\·\·& t_1+{\mathrm{\τ}}_{1N}\\ t_2+{\mathrm{\τ}}_{21}& t_2+{\mathrm{\τ}}_{22}& \·\·\·& t_2+{\mathrm{\τ}}_{2N}\\ t_3+{\mathrm{\τ}}_{31}& t_3+{\mathrm{\τ}}_{32}& \·\·\·& t_2+{\mathrm{\τ}}_{3N}\end{array}\right)---(1-8)$

Definition binary variable θ _ifor:

Therefore the target of the different x time optimization problem with reference to message is for maximizing θ problem

$\mathrm{\θ}=\underset{T}{\max }\left({\mathrm{\Σ}}_{i=1}^N{\mathrm{\θ}}_i\right)---(1-9)$

Suppose to represent that 3 arrive by the mistiming of location node with reference to message, then can obtain formula (1-10) as follows with λ:

$\left\{\begin{array}{c}{\mathrm{\λ}}_{1i}=|t_1+{\mathrm{\τ}}_{1i}-t_2-{\mathrm{\τ}}_{2i}|\\ {\mathrm{\λ}}_{2i}=|t_1+{\mathrm{\τ}}_{1i}-t_3-{\mathrm{\τ}}_{3i}\\ {\mathrm{\λ}}_{3i}=|t_1+{\mathrm{\τ}}_{2i}-t_3-{\mathrm{\τ}}_{3i}|\end{array},|\right.---(1-10)$

Due to the transmission time be necessary on the occasion of, then can obtain following two constraint conditions, as formula (1-11) and formula (1-12) from formula (1-9), formula (1-10):

$1,\left\{\begin{array}{c}{t}_1\≥0\\ t_2\≥0\\ t_3\≥0\end{array}\right.---(1-11)$

$2,{\mathrm{\θ}}_i\≤1-\frac{{\mathrm{\λ}}_{\mathrm{mi}}-\mathrm{\δ}}{M},m=\mathrm{1,2,3}---(1-12)$

M wherein in formula (1-12) is large positive integer, formula (1-9), formula (1-10), formula (1-11) and formula (1-12) be can be regarded as: if do not receive all reference messages in time period δ, then λ is greater than δ to be a little positive number, then just hook, therefore want constraint formulations (1-11) and formula (1-12) to set up, θ must be 0, and the condition that this and θ are defined as O meets.

Equally, if having received all Reference News in time period δ, then to be negative, for just, therefore constraint formulations (1-11) and formula (1-12) will be made to set up, θ can be 0 or 1, and due to be do maximize process, then θ tends to 1, also the condition being defined as 1 with θ meets, and therefore proves that constraint formulations (1-11) and formula (1-12) are correct and significant, therefore can as constraint condition.

Its optimization problem can be converted into linear optimization problem by said method, available standards linear optimization algorithm such as Branch-and-Cut solves.This optimization problem determines the transmission time of each Reference News, and decreases ambulant impact in transmitting procedure, effectively can improve positioning precision.

Claims (4)

1. one kind adapts to ambulant water sound sensor network localization method, it is characterized in that: propose to utilize maximization in stipulated time section, to be received the probability of the reference message of all different reference modes to optimize the method in the reference message transmission moment of different reference mode by location node, it can adapt to the intrinsic movability of node preferably, improves positioning precision.

2. the ambulant water sound sensor network localization method of adaptation as claimed in claim 1, the present invention proposes by calculating the positioning error that different reference mode set produces, the method of the optimum reference mode set that regioselective error is minimum, it can adapt to the intrinsic movability of node preferably, improves positioning precision.

3. the ambulant water sound sensor network localization method of adaptation as claimed in claim 1, The present invention gives a kind of mathematical modeling scheme optimizing the reference message transmission moment of different reference mode, the reference message transmission moment of different reference mode can be got by this mathematical modeling scheme, reduce the intrinsic ambulant impact of node, improve positioning precision.

4. the ambulant underwater sound sensing network localization method of adaptation as claimed in claim 2, The present invention gives a kind of by calculating the positioning error that different reference mode set produces, select the mathematical modeling scheme of optimum reference mode set, optimum reference mode set can be got by this mathematical modeling scheme, improve positioning precision.