阅读说明：本技术 一种基于距离-方位积分字典的目标定位方法 (Target positioning method based on distance-orientation integral dictionary ) 是由 杨长生 李杭波 梁红 于 2021-07-12 设计创作，主要内容包括：本发明涉及一种基于距离-方位积分字典的目标定位方法,针对不同环境建立不同的距离-方位联合积分字典,从而可以对目标的距离及方位进行联合估计。因而可以实现在不同环境下均可对目标进行较好的距离-方位联合估计,且误差较小,在常规距离-方位字典方法精度不变的情况下大量减少了运算时间。有益效果在于：该定位方法运算量小,精度高,可以针对不同环境,在不同干扰下建立不同的字典,从而可以满足在各种环境条件下均可准确的定位目标距离位置,经试验验证,达到了非常好的效果。能够有效解决传统方法中的环境干扰对测量影响的问题,具有广泛的应用前景,可直接投入使用。(The invention relates to a target positioning method based on a distance-orientation integral dictionary, which aims at establishing different distance-orientation combined integral dictionaries aiming at different environments so as to carry out combined estimation on the distance and the orientation of a target. Therefore, better distance-direction joint estimation can be carried out on the target under different environments, the error is smaller, and the operation time is greatly reduced under the condition that the precision of the conventional distance-direction dictionary method is not changed. Has the advantages that: the positioning method has small computation amount and high precision, and can establish different dictionaries under different interferences aiming at different environments, thereby being capable of accurately positioning the distance position of the target under various environmental conditions, and achieving very good effect through experimental verification. The method can effectively solve the problem that the environmental interference in the traditional method influences the measurement, has wide application prospect and can be directly put into use.)

1. A target positioning method based on a distance-orientation integral dictionary is characterized in that: l array elements form a planar array, and a space domain is divided intoEach azimuth theta_j(j＝1,2,…,N_s) All corresponding to a potential target source signal r_i(i ═ 1,2, …, n); the method comprises the following steps:

step 1, constructing a distance-orientation combined dictionary by using the received target echo:

the mth column of the distance-orientation joint dictionary is represented as:

wherein m is (Lxn) (j-1) + k, j is more than or equal to 1 and less than or equal to N_s，1≤k≤L×n，1≤m≤L×n×N_s；

Wherein s is_l(r_i,θ_j) For the first array element at a distance r_iDirection is theta_jA target echo received;

step 2, integrating the distance-azimuth dictionary (1) to construct a first-level distance-azimuth integration dictionary B;

firstly, dividing the whole distance interval and the whole azimuth interval into Q small intervals, wherein Q is M N, M is the number of the divided distance intervals, and N is the number of the divided azimuth intervals;

the signal is then integrated over each interval:

wherein theta is_qAnd theta_q-1And r_qAnd r_q-1Upper and lower limits of the qth region respectively

Obtaining a new distance and azimuth integral dictionary:

B＝[b₁ b₂ ... b_Q]

wherein: b_q＝[b_q(t₁) b_q(t₂) ... b_q(t_N)]′

The dictionary is specifically represented as:

the vector with subscripts of M and N represents atoms formed by the area integrals corresponding to the mth distance interval and the nth azimuth interval, and the number of dictionary atoms after distance and azimuth integrals is Q-M-N;

and step 3: and (3) representing the received echoes of the target to be detected as the linear superposition of the echoes received by the distance-orientation integral dictionary at different distances and orientations in the step 2:

E(t)＝∑_i,je(t_i,θ_j)＝∑_i,jA_ijb_q(r_i,θ_j)

wherein r is_iDenotes the distance, θ_jIndicates the orientation, A_ijRepresenting the amplitude of the echo, b_q(r_i,θ_j) Represents the distance r_iAzimuth theta_jNormalized echo of (1);

and 4, step 4: receiving a new echo E of a target in a certain scene, and sparsely expressing the new echo E in a constructed first-level distance and azimuth integration dictionary B, wherein the target echo is expressed as:

wherein R is sparse expression of the echo E in a first-level distance and azimuth integration dictionary B;

and 5: and (3) introducing an L1 norm minimization method in a convex optimization theory to solve, and roughly estimating the distance and the direction of the target by using a first-level distance-direction integral dictionary:

wherein, σ is the variance of noise, p is the potential of a dictionary, and γ represents a weight coefficient, and the magnitude increases along with the enhancement of noise; the sparse vector R is obtained, and then the preliminary m is obtained₁And n₁Approximately determine r_mAnd theta_n。

Step 6: finding m from a first-order distance-orientation integral dictionary₁And n₁Constructing an original distance and orientation dictionary D with the atomic distance of 10 multiplied by 10 as a center, wherein the total number of dictionary atoms is 100, and positioning the target in the area;

and 7: and (3) sparsely expressing the newly received echo E in the step (4) in the constructed secondary distance and orientation dictionary D again, and expressing the target echo as follows:

E＝αD

wherein alpha is sparse representation of the echo E in the secondary distance and orientation dictionary D;

and 8: introducing an L1 norm minimization method in a convex optimization theory to solve, and utilizing a secondary distance-orientation dictionary E to carry out distance r on the target_mAnd the orientation theta_nAnd (3) carrying out accurate estimation:

wherein, σ is the variance of noise, p is the potential of a dictionary, and γ represents a weight coefficient, and the magnitude increases along with the enhancement of noise; wherein alpha is sparse expression of echo E in a secondary distance and orientation dictionary D, corresponding m and n are obtained, and then r is obtained_mAnd theta_nThe specific azimuth distance of the target.

Technical Field

The invention belongs to the field of target parameter estimation, and relates to a target positioning method based on a distance-azimuth integration dictionary.

Background

At present, more traditional target parameter estimation methods include methods utilizing preformed beam orientation, split beam orientation, interpolation method orientation, multi-beam orientation and the like. The traditional target parameter estimation method is based on the principle that the solution is carried out after the acoustic path difference or the phase difference between array elements of a target signal reaching a matrix is measured. The theoretical research and the practical application of the method are mature, but the method has many defects, such as obvious precision reduction under severe conditions, higher precision requirement of measuring equipment, poor fault tolerance, larger error influence of the environment, and limited application occasions. In order to improve the target measurement accuracy under different environments, the measurement method is simpler to operate and wider in application, and the invention is very necessary.

Maksim Butsenko et al, in the document "Maksim, Butsenko, Johan, et al, estimating space Signals Using Integrated Wideband dictionary [ J ]. IEEE Transactions on Signal Processing, 2018", propose a Sparse expression parameter estimation method based on a time delay one-dimensional integral dictionary, construct a Wideband dictionary by integrating dictionary elements, thereby reducing the size of the Sparse expression dictionary, which can reduce the overall computation amount, but only for one-dimensional dictionary integration, and does not extend to two-dimensional.

Disclosure of Invention

Technical problem to be solved

In order to avoid the defects of the prior art, the invention provides a target positioning method based on a distance-orientation integral dictionary.

Technical scheme

A target positioning method based on a distance-orientation integral dictionary is characterized in that: l array elements form a planar array, and a space domain is divided intoEach azimuth theta_j(j＝1，2，…，N_s) All corresponding to a potential target source signal r_i(i ═ 1,2, …, n); the method comprises the following steps:

step 1, constructing a distance-orientation combined dictionary by using the received target echo:

the mth column of the distance-orientation joint dictionary is represented as:

wherein m is (Lxn) (j-1) + k, and L is less than or equal to j and less than or equal to N_s，1≤k≤L×n，1≤m≤L×n×N_s；

Wherein s is_l(r_i，θ_j) For the first array element at a distance r_iDirection is theta_jA target echo received;

step 2, integrating the distance-azimuth dictionary (1) to construct a first-level distance-azimuth integration dictionary B;

firstly, dividing the whole distance interval and the whole azimuth interval into Q small intervals, wherein Q is M N, M is the number of the divided distance intervals, and N is the number of the divided azimuth intervals;

the signal is then integrated over each interval:

wherein theta is_qAnd theta_q-1And r_qAnd r_q-1Upper and lower limits of the qth region respectively

Obtaining a new distance and azimuth integral dictionary:

B＝[b₁ b₂...b_Q]

wherein: b_q＝[b_q(t₁)b_q(t₂)...b_q(t_N)]′

The dictionary is specifically represented as:

the vector with subscripts of M and N represents atoms formed by the area integrals corresponding to the mth distance interval and the nth azimuth interval, and the number of dictionary atoms after distance and azimuth integrals is Q-M-N;

and step 3: and (3) representing the received echoes of the target to be detected as the linear superposition of the echoes received by the distance-orientation integral dictionary at different distances and orientations in the step 2:

E(t)＝∑_i，je(t_i，θ_j)＝∑_i，jA_ijb_q(r_i，θ_j)

wherein r is_iDenotes the distance, θ_jIndicates the orientation, A_ijRepresenting the amplitude of the echo, b_q(r_i，θ_j) Represents the distance r_iAzimuth theta_jNormalized echo of (1);

and 4, step 4: receiving a new echo E of a target in a certain scene, and sparsely expressing the new echo E in a constructed first-level distance and azimuth integration dictionary B, wherein the target echo is expressed as:

E＝R

wherein R is sparse expression of the echo E in a first-level distance and azimuth integration dictionary B;

and 5: and (3) introducing an L1 norm minimization method in a convex optimization theory to solve, and roughly estimating the distance and the direction of the target by using a first-level distance-direction integral dictionary:

wherein, σ is the variance of noise, p is the potential of a dictionary, and γ represents a weight coefficient, and the magnitude increases along with the enhancement of noise; the sparse vector R is obtained, and then the preliminary m is obtained₁And n₁Approximately determine r_mAnd theta_n。

Step 6: finding m from a first-order distance-orientation integral dictionary₁And n₁Constructing an original distance and orientation dictionary D with the atomic distance of 10 multiplied by 10 as a center, wherein the total number of dictionary atoms is 100, and positioning the target in the area;

and 7: and (3) sparsely expressing the newly received echo E in the step (4) in the constructed secondary distance and orientation dictionary D again, and expressing the target echo as follows:

E＝αD

wherein alpha is sparse representation of the echo E in the secondary distance and orientation dictionary D;

and 8: introducing an L1 norm minimization method in a convex optimization theory to solve, and utilizing a secondary distance-orientation dictionary E to carry out distance r on the target_mAnd the orientation theta_nAnd (3) carrying out accurate estimation:

wherein, σ is the variance of noise, p is the potential of a dictionary, and γ represents a weight coefficient, and the magnitude increases along with the enhancement of noise; wherein alpha is sparse expression of echo E in a secondary distance and orientation dictionary D, corresponding m and n are obtained, and then r is obtained_mAnd theta_nThe specific azimuth distance of the target.

Advantageous effects

The invention provides a target positioning method based on a distance-orientation integral dictionary, which is used for establishing different distance-orientation combined integral dictionaries aiming at different environments, so that the distance and the orientation of a target can be jointly estimated. Therefore, better distance-direction joint estimation can be carried out on the target under different environments, the error is smaller, and the operation time is greatly reduced under the condition that the precision of the conventional distance-direction dictionary method is not changed.

The invention has the beneficial effects that: the positioning method has small computation amount and high precision, and can establish different dictionaries under different interferences aiming at different environments, thereby being capable of accurately positioning the distance position of the target under various environmental conditions, and achieving very good effect through experimental verification. The method can effectively solve the problem that the environmental interference in the traditional method influences the measurement, has wide application prospect and can be directly put into use.

The dictionaries established in different environments can be repeatedly used, measurement is carried out for multiple times, and different dictionaries are used in different environments, so that the calculated amount is minimum, and the resource utilization is maximum.

Drawings

FIG. 1: is a flow chart of the invention.

FIG. 2; common distance-orientation joint dictionary

(a) Normalized distance-azimuth dictionary (b), distance-azimuth dictionary coherence characteristic diagram (c), coherence characteristic contour diagram

FIG. 3: common distance-orientation joint dictionary estimation result

(a)SNR＝0dB (b)SNR＝-10dB (c)SNR＝-20dB

FIG. 4: first-order distance-azimuth joint integral dictionary

(a) Normalized distance-azimuth dictionary (b), distance-azimuth dictionary coherence characteristic diagram (c), coherence characteristic contour diagram

FIG. 5: first-order distance-orientation joint integral dictionary estimation result

(a)SNR＝0dB (b)SNR＝-10dB (c)SNR＝-20dB

FIG. 6: second-order distance-azimuth joint integral dictionary estimation result

(a)SNR＝0dB (b)SNR＝-10dB (c)SNR＝-20dB

Fig. 2 is a distance-azimuth joint integration and coherence analysis thereof established by the embodiment of the invention, wherein a general distance-azimuth joint dictionary, dictionary coherence characteristics, and contour map are respectively shown from top to bottom. Distance-azimuth joint integral dictionary, dictionary coherence properties, contour map.

FIG. 3 is a diagram of the estimation results of the invention at different SNR.

Detailed Description

The invention will now be further described with reference to the following examples and drawings:

the basic idea of the invention is to establish a distance-azimuth joint dictionary, perform integration processing on the distance-azimuth joint dictionary, establish a distance-azimuth joint integration dictionary, firstly determine a rough region according to a received target echo by applying a convex optimization theory, establish a fine dictionary in the region, and estimate a specific distance and azimuth of a target.

The technical scheme adopted by the invention for solving the technical problems is as follows: a solving algorithm for estimating the range and the azimuth of a target by processing the echo of the target by establishing a range-azimuth joint integral dictionary mainly comprises the following steps:

1) and receiving target echoes in different directions by using the planar array. Assuming that a planar array is composed of L array elements, the spatial domain is divided into L array elements without considering the influence of a pitch angleAnd assuming each possible orientation theta_j(j＝1，2，…，N_s) All corresponding to a potential target source signal r_iAnd (i is 1,2, …, n), and target echoes in different directions are received by using a planar array.

Specifically, the method comprises the following steps: and reading target echoes in all directions, and finishing the target echoes in a silencing water pool, wherein the transducer is positioned at 2m under water, and the distance between the transmitting array element and the center of the receiving array is 2.1 m. The angle range of the rotation of the receiving array is-30 degrees to 30 degrees, and the step length is 1 degree.

2) And constructing a distance-direction joint dictionary by using the received target echoes.

A combined range-azimuth dictionary is built from the received data and the received echo signals e (t) are represented as a linear superposition of the received echoes at different ranges and azimuths. The mth column of the distance-orientation joint dictionary may be represented as:

wherein m is (Lxn) (j-1) + k, j is more than or equal to 1 and less than or equal to N_s，1≤k≤L×n，1≤m≤L×n×N_s。

Wherein s is_l(r_i，θ_j) For the first array element at a distance r_iDirection is theta_jA target echo received;

the array aperture is 0.9m, and the distance between the transmitting array element and the middle point of the two array elements is 2.1 m; in the azimuth dimension, 61 atoms, the range of azimuth angles is-30 degrees to 30 degrees, and the step length is 1 degree; in the distance dimension, 100 atoms are used at each angle, the distance ranges from 2.1m to 5.1m, the relative delay between two adjacent atoms is 10 sample delay points, and finally, a distance-orientation joint dictionary as shown in the attached figure 2 of the specification and coherent characteristic analysis are obtained.

As can be seen from FIG. 2, the distance-orientation joint dictionary is jagged, which is determined by the construction method of the dictionary; the values of the coherence of the distance-azimuth combined dictionary on the diagonal are all 1, the other values are small, the width of the main lobe is narrow, and the height of the side lobe is also low; it is explained that the difference between atoms in the distance-orientation dictionary is large, and the estimation precision is higher when the distance and the orientation of the target are estimated.

3) And constructing a primary distance-orientation integral dictionary. And integrating the distance-orientation dictionary, namely dividing the whole distance interval and the orientation interval into Q small intervals, wherein Q is M N, M is the number of the divided distance intervals, and N is the number of the divided orientation intervals.

Table 1 shows the simulation result, and it can be seen that, when the integration interval is 1, the constructed dictionary is a common dictionary defined on the grid, and as the integration interval increases, the distance error and the orientation error both increase. The range error does not vary much and the azimuth error is more sensitive to the integration interval. In order to reduce estimation error, and also effectively reduce operation time, in the following simulation, we select an integral dictionary with M being 3 and N being 3.

Table 1: integral dictionary estimation error

The signal is integrated over each interval to obtain a new type of dictionary element. In this definition method, the dictionary matrix no longer covers only certain parameter points, but the entire continuous parameter interval.

Wherein theta is_qAnd theta_q-1And r_qAnd r_q-1The upper and lower limits of the qth region, respectively, the integral dictionary can be expressed as:

B＝[b₁ b₂...b_Q]

wherein

b_q＝[b_q(t₁)b_q(t₂)...b_q(t_N)]′

The dictionary is specifically represented as:

and the vectors with subscripts of M and N represent atoms formed by the integration of the regions corresponding to the mth distance interval and the nth azimuth interval, and the number of the dictionary atoms after the integration is Q-M-N. The dictionary matrix can be applied to solve the L1 convex optimization algorithm. In the definition of the initial dictionary, the number of elements can be less than that of the conventional dictionary, and the integral dictionary is defined on the region, so that the risk of mismatch can be effectively reduced. Of course, after the initial parameter range is obtained, the area which is divided again is more refined, the range of the divided area is small, in this case, the integral dictionary is similar to the conventional dictionary, especially when the integral area is extremely narrow, the integral dictionary is about equivalent to the conventional dictionary, and in this case, the conventional dictionary can be used for solving. Thus, the primary purpose of the integrated dictionary is to replace the initial dictionary, whereas a conventional dictionary may be used in a subsequent refined dictionary.

The number of dictionary atoms after the specific integration is Q33 × 20. Finally, a first-order distance-orientation integral dictionary as shown in the specification and attached figure 3 is obtained, and the coherence characteristic analysis is carried out.

4) By receiving the echo of the target to be measured, it is expressed as 3) the linear superposition of the echo received by the distance-orientation integral dictionary at different distances and orientations:

wherein r is_iDenotes the distance, θ_jIndicates the orientation, A_ijRepresenting the amplitude of the echo, b_q(r_i，θ_j) Represents the distance r_iAzimuth theta_jNormalized echo of (1);

5) receiving target echo, and sparsely expressing the target echo in a first-order integral dictionary

E＝RB

Wherein, R is the sparse expression of the echo in the first-order integration dictionary.

6) And roughly estimating the distance and the direction of the target by using a primary distance-direction integral dictionary. Due to condition limitation, 5) cannot completely and accurately solve the distance and orientation information of the target, so that an L1 norm minimization method in convex optimization theory is introduced to solve:

wherein, σ is the variance of noise, p is the potential of a dictionary, and γ represents a weight coefficient, and the magnitude increases with the enhancement of noise.

The method specifically comprises the following steps:

the value range of gamma is more than 0 and less than gamma_max＝||B^Tx||_∞And an empirical formula:

where σ is the noise variance and p is the dictionary potential.

The distance and orientation information of the target can be solved by an L1 norm minimization method in a convex optimization theory, namely:

the final sparse vector R is solved to obtain approximate distance and orientation information of the target, and the result is shown in fig. 4, and the approximate distance and orientation information is used as the basis for constructing the secondary dictionary D. The algorithm steps are similar to step four.

The specific form of the secondary dictionary D is as follows:

the target echo can be expressed as

E＝αD

Wherein α is a sparse representation of the echo in the secondary dictionary, and the specific distance and orientation information of the target can be obtained according to the result, which is shown in fig. 5.

Under the condition that the signal-to-noise ratio is 0dB, -10dB, -20dB, the dictionary is used for completing the joint estimation of the distance and the direction of two targets. Assuming that the target position is (0 °,2.67m), the results of the joint estimation of the target distance and the target position are shown in fig. 3, 5 and 6 under the conditions that the signal-to-noise ratio is 0dB, -10dB and-20 dB, respectively. Fig. 3 shows the common distance-orientation dictionary estimation result, and fig. 5 and 6 show the distance-orientation dictionary integral and two-level dictionary estimation result.

It can be seen that, when the data is higher than-20 dB, the estimation error of the common dictionary is the same as that of the second-level distance-azimuth joint integral dictionary in estimation precision, and the precision is higher than that of the first-level distance-azimuth joint integral dictionary; and the second-level distance and azimuth joint integral dictionary has similar operation time with the first-level distance and azimuth joint integral dictionary, and both the operation time and the operation time are far smaller than those of the original common distance and azimuth joint dictionary. Table 2 compares the distance-orientation joint integral dictionary presented herein with a conventional distance-orientation dictionary.

Table 2: dictionary estimation error and estimation time

It can be concluded that under low signal-to-noise ratio, a two-level distance-orientation joint integration dictionary can be used to replace a common distance-orientation joint dictionary, and the integration dictionary is obviously higher than a conventional dictionary in computational efficiency. Because the area covered by each element in the integral dictionary is wider, the problem of mismatch caused by too large interval of the conventional dictionary is avoided, the calculation time is reduced, and the calculation accuracy is the same as that of the common distance-orientation combined dictionary in low signal-to-noise ratio.