Distributed array structure optimization method and system of quantum fox mechanism, and storage medium

Abstract

The invention discloses a distributed array structure optimization method and system of a quantum fox mechanism and a storage medium, and the method comprises the steps: firstly carrying out the modeling of space impact noise and information source information, constructing an infinite norm weighted fraction low-order moment matrix through snapshot data, and carrying out the direction estimation, and obtaining a root-mean-square error equation; initializing a quantum fox group, and determining local and global optimal measurement positions; respectively updating the quantum rotation angle and position of the fox by adopting a development strategy and an exploration strategy so as to obtain a measurement position, substituting the measurement position into a fitness function, and updating an optimal position by judging a fitness value; judging whether the maximum number of iterations is reached or not, and if the maximum number of iterations is reached, outputting a global optimal measurement position and converting the global optimal measurement position into array layout structure information; and if not, continuing iteration. A designed development and exploration two-stage mechanism is improved by means of a quantum optimization theory, and quantum coding and evolution are introduced, so that local optimum can be jumped out, convergence can be improved, noise suppression capability can be enhanced during signal processing, and effectiveness and superiority of array layout can be guaranteed.

Description

Technical Field

[0001] The present invention belongs to the field of array signal processing, and particularly relates to an optimization method, system and storage medium for a distributed array structure with a quantum fox mechanism. Background Art

[0002] In array signal processing, distributed arrays, as an important branch, have been widely used, such as high-speed transmission, signal parameter estimation, and direction-of-arrival estimation. Compared with a single sensor, a distributed sensor array can cover a wider area. By integrating the measurement results of multiple sensors, it can effectively avoid the errors that may exist in a single sensor, and is applicable to more specific application scenarios to improve the selectivity and reliability of the system. However, the current distributed array system often relies on engineering experience for the position arrangement of sub-arrays, such as equal-spacing arrangement or uniform arrangement. Different array arrangement methods will affect the direction-finding accuracy and performance of the distributed array, and for different application scenarios, it is often necessary to repeatedly adjust the element spacing according to experience to obtain a better direction-finding effect. Especially in the case of small snapshot scenarios, the traditional sub-array position arrangement method may cause the performance of the distributed array to deteriorate or even fail. Therefore, it is of great significance and value to study and design a high-performance distributed array layout structure for an impact noise environment and a small number of sampling snapshots.

[0003] Through the retrieval of existing technical literature, it is found that Li J, Zhang Q, Deng W, etc. proposed an algorithm for source direction finding and direct localization using a UAV array in the paper "Source Direction Finding and Direct Localization Exploiting UAV Array With Unknown Gain-Phase Errors" published in the 《IEEE Internet of Things Journal》. By constructing a cost function and obtaining a spectral function, the direction and position of the source can be estimated using grid search, and the gain-phase error value can be continuously solved. Its experimental environment is in a Gaussian noise environment, and whether the direction finding performance is good under harsh conditions remains to be studied. Wang Xu et al. proposed an innovative grating lobe suppression algorithm in the paper "A Sparse Uniform Array Grating Lobe Suppression Method Using Dual-Carrier Pattern Multiplication" published in the 《Journal of Electronics & Information Technology》, aiming at the characteristic that there are differences in the relative positions of the main lobe and the grating lobe in the carrier frequency array pattern. This algorithm quantitatively analyzes the key parameters affecting the grating lobe suppression performance and deduces the mathematical relationship between the peak sidelobe ratio and the frequency difference after grating lobe suppression. By effectively using the echo information at different carrier frequencies, this algorithm avoids a large-scale search process and significantly reduces the computational complexity. However, its main goal is to optimize the sidelobe level, and further research is needed for the optimization of the actual direction finding performance. Chen M, Mao X et al. proposed a deep neural network framework for DOA estimation with a feasible computational complexity in the paper "Direction of arrival estimation of unknown emitter by deep neural networks with array imperfections" published in the 《IET Radar, Sonar & Navigation》. The two-stage algorithm consists of a detection network and a series of parallel DOA estimation networks, which has potential usefulness for improving the accuracy of mutual coupling DOA estimation without simplifying the mutual coupling matrix. However, this method can achieve high accuracy when the number of snapshots is large, but the performance deteriorates when the number of snapshots for small blocks is sampled, and it is difficult to achieve the expected effect.Zhang YD, Chowdhury MW et al. in "Direction-of-arrival estimation in closely distributed array exploiting mixed-precision covariance matrices" published in *Signal Processing* use multiple closely spaced subarrays to maximize the potential of the distributed array while minimizing the communication overhead between the subarrays and the processing center, and combine it with the full-precision subarray autocovariance matrix to create a mixed-precision covariance matrix for the entire array. However, the proposed distributed array structure in this method is fixed. It has a high estimation accuracy for dealing with specific environments, but the direction-of-arrival estimation results in harsh environments such as small snapshot environments need to be further studied.

[0004] The existing literature search results show that the existing distributed array optimization methods mainly focus on optimizing the sidelobe level and the covariance matrix of the subarrays, which greatly improve the estimation performance and accuracy for large sampling snapshots in a Gaussian noise environment. However, the performance deteriorates when estimating in a small sampling snapshot environment, and there is a lack of research on the performance in harsh environments such as impulse noise. Therefore, this method directly focuses on the direction-of-arrival estimation performance of the distributed array in an impulse noise and small sampling snapshot environment. By processing the received signal with the infinite norm weighted fractional lower-order moment, its noise suppression ability is improved. The root mean square error of the source estimation of the distributed array in practical applications is used as the optimization objective to obtain the optimal distributed array structure and improve the direction-finding performance. At the same time, to reduce the computational complexity, a solution combining an intelligent optimization algorithm and a quantum algorithm is proposed to further improve the effectiveness of the algorithm. Therefore, it is of great significance and practical value to design a distributed array structure based on the quantum fox optimization mechanism to solve the problem in an impulse noise and small sampling snapshot environment. Summary of the Invention

[0005] The purpose of the present invention is to provide an optimization method, system and storage medium for a distributed array structure based on the quantum fox mechanism, which can generate a more adaptable distributed array structure in different environments, thereby further ensuring the effectiveness and superiority of the optimized subarray layout.

[0006] The purpose of the present invention is achieved through the following technical solutions:

[0007] An optimization method for a distributed array structure based on the quantum fox mechanism, comprising the following steps:

[0008] Step 1: Model the impulse noise and source information in space;

[0009] Step 2: Use the snapshot data received by the array to construct an infinite norm weighted fractional lower order moment matrix for direction estimation, and obtain the root mean square error equation of direction estimation;

[0010] Step 3: Initialize the quantum fox swarm to obtain the local optimal measurement position and the global optimal measurement position;

[0011] Step 4: Update the quantum rotation angle and quantum position of each fox using the exploitation strategy and obtain the measurement position;

[0012] Step 5: Update the quantum rotation angle and quantum position of each fox using the exploration strategy and obtain the measurement position;

[0013] Step 6: Substitute the measurement position obtained by the nth fox in the (t + 1)th generation into the fitness function to obtain the fitness value, and judge the fitness value to obtain the local optimal measurement position and the global optimal measurement position;

[0014] Step 7: Judge whether the maximum number of iterations is reached. If so, output the global optimal measurement position and convert it into the array layout structure information; otherwise, let t = t + 1 and return to Step 4 to continue the iteration.

[0015] Further, the specific content of Step 1 is as follows:

[0016] Set the noise in the space to be impulse noise obeying the α-stable distribution, with the characteristic exponent α and the symmetry parameter β; set K known narrowband signal sources with wavelength λ in the space, where the azimuth angles θ = [θ1, θ2,... θ K , -π ≤ θ k ≤ π; assume a distributed array contains M s sub-arrays, each sub-array has M z array elements, then the total number of array elements of the entire array is M = M z × M s , the element spacing within the sub-array is d, the arrangement of the sub-arrays in the space is one-dimensional, and the spacing between each sub-array is an integer multiple of the element spacing.

[0017] Further, the specific content of Step 2 is as follows:

[0018] Let the qth snapshot data received be y(q) = A(θ)s(q) + n(q), where q is the snapshot number label and q ∈ [1, Q], Q is the maximum number of snapshots, A(θ) is the steering vector matrix, s(q) is the signal vector, and n(q) is the independent and identically distributed impulse noise vector; perform infinite norm normalization processing on the snapshot data where, The infinite norm weighted fractional lower order moment matrix of the qth snapshot data Among them, \(C(q)\) is the fractional lower-order moment matrix of the \(q\)-th snapshot data, \(C(q)=[c_1(q),c_2(q),\cdots,c M (q)]\) is an \(M\times M\) matrix, where \(c n (q)=[c 1n (q),c 2n (q),\cdots,c Mn (q)] T , \(n = 1,2,\cdots,M\), and the element in its row and column is Among them, \(p_0\) is the lower-order moment parameter, and \((\cdot) * \) represents taking the conjugate; the processed infinite norm weighted fractional lower-order moment matrix is used as the covariance matrix, and then the MUSIC algorithm is used for DOA estimation. After the covariance matrix \(R\) is eigen-decomposed, it is divided into two subspaces: the signal subspace \(U S \) and the noise subspace \(U N , The signal subspace \(U S \) is composed of the eigenvectors corresponding to the larger eigenvalues, and these eigenvalues form a large eigenvalue diagonal matrix \(\sum S \); while the noise subspace \(U N \) is composed of the remaining eigenvectors, which correspond to the smaller eigenvalues, and these eigenvalues form another small eigenvalue diagonal matrix \(\sum N \); when the noise subspace \(U N \) is right-multiplied by the covariance matrix \(R\), we get Since the covariance matrix \(R S \) is full-rank and invertible, through derivation, it is obtained that the steering vector \(A(\theta)\) is orthogonal to the noise subspace \(U N , that is, \(A H (\theta)U N = 0\); based on this orthogonality, the spatial spectrum function is further derived Then, the estimated azimuth value is obtained by searching for the spectral peak in the th experiment Substitute the direction finding result into the objective function Among them, is the unknown to be solved for the array position structure, \(K\) is the number of known signal sources in space, \(\theta k is the accurate azimuth value of the \(k\)-th signal source, is the azimuth estimation value of the \(k\)-th signal source in the th experiment at the array position \(z\).

[0019] Furthermore, the specific content of step three is:

[0020] First, assume that there are \(N\) quantum foxes in space, and each quantum fox is evenly distributed. Then, the quantum position of the \(n\)-th quantum fox in the \(t\)-th generation is The measurement equation is to obtain the measurement position The corresponding encoded array position is where is a random number uniformly distributed between [0, 1], O is the maximum dimension of the solution space, and the maximum number of iterations is set to T; the array position mapped by the position of the nth fox obtained is brought into the fitness function for evaluation, The fitness function of is calculated to obtain the fitness value of the position of the nth fox, where θ is the accurate azimuth value of the kth signal source, k is the estimated direction of the kth incoming wave obtained after executing the direction finding algorithm at the array position of the nth fox in the tth generation and the tth experiment, and then the optimal measurement position of the nth fox up to the tth generation is recorded as the local optimal position The measurement position with the optimal fitness up to the tth generation is recorded as the global optimal position

[0021]

[0022] Furthermore, the specific content of step four is as follows:

[0022] For the nth fox, set the hunting opportunity probability as a random number uniformly distributed between [0, 1]. If is greater than γ1, the fox moves towards the global optimal solution, and calculate the sound propagation distance where the sound propagation time is defined as a random number distributed between [0, 1], and find the sound speed through the best position Then the distance between the fox and the prey is half of the distance found through the sound wave After finding the distance between the fox and the prey, find the new position through the jump height. The average time of the nth fox in the tth generation is is a random number uniformly distributed between [0, 1], and the jump height is Define the oth - dimensional quantum rotation angle of the nth quantum fox in the (t + 1)th generation as If is less than γ1, the oth - dimensional quantum rotation angle of the nth quantum fox in the (t + 1)th generation is defined as where is a random number uniformly distributed between [0, γ1], is a random number uniformly distributed between [γ1, 1]; the evolution method of the oth - dimensional quantum position of the nth quantum fox is where is a random number uniformly distributed in [0, 1], b1 represents the probability of performing an inversion operation on this qubit when the quantum rotation angle is 0, and its value is a constant in [0, 1 / O]; finally, the corresponding position is obtained by measuring the quantum position Its measurement equation is where is a random number uniformly distributed in [0, 1], n = 1, 2,..., N, o = 1, 2,..., O; substituting the measurement position of the nth fox into the fitness function to obtain the fitness value If then the local optimal measurement position of the nth fox until the (t + 1)-th generation is Otherwise,[[]] select the optimal position from the N local optimal measurement positions until the (t + 1)-th generation and set it as the global optimal measurement position until the (t + 1)-th generation

[0023] Furthermore, the specific content of the fifth step is as follows:

[0024] In the exploration stage, the fox hunts by random walking. To ensure that the fox moves towards the optimal position, a small variable and variable a are used to control the search, is the minimum average time of the nth individual up to the t-th generation, is a random number uniformly distributed in [0, 1], is the average time of the nth individual in the t-th generation; the random variable where T is the maximum number of iterations; define the o-th dimensional quantum rotation angle of the nth quantum fox in the (t + 1)-th generation as where is a random number uniformly distributed in [0, 1], and then define the evolution method of the o-th dimensional quantum position of the nth quantum fox as where is a random number uniformly distributed in [0, 1], b1 represents the probability of performing an inversion operation on this qubit when the quantum rotation angle is 0, and its value is a constant in [0, 1 / O]; finally, the corresponding position is obtained by measuring the quantum position, and its measurement equation is where is a random number uniformly distributed in [0, 1], n = 1, 2,..., N, o = 1, 2,..., O.

[0025] Furthermore, the specific content of the sixth step is as follows:

[0026] The measured position obtained by the nth fox in the (t + 1) - th generation is substituted into the fitness function to obtain the fitness value If then determine the local optimal measurement position up to the (t + 1) - th generation Otherwise the value remains unchanged; select the optimal position from the N local optimal measurement positions up to the (t + 1) - th generation and set it as the global optimal measurement position up to the (t + 1) - th generation

[0027] A computer device / apparatus / system includes a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the steps of an optimization method for the distributed array structure of a quantum fox mechanism.

[0028] A computer - readable storage medium stores a computer program / instructions. When the computer program / instructions are executed by a processor, the steps of an optimization method for the distributed array structure of a quantum fox mechanism are implemented.

[0029] A computer program product includes a computer program / instructions. When the computer program / instructions are executed by a processor, the steps of an optimization method for the distributed array structure of a quantum fox mechanism are implemented.

[0030] The beneficial effects of the present invention are as follows:

[0031] 1. The present invention proposes an optimization method for the array structure based on the quantum fox optimization mechanism. Different from the traditional method that takes the grating lobe suppression of the array pattern as the optimization goal in the Gaussian noise environment, this method simulates the small - block beats that may appear in the actual direction - finding results, generates the corresponding array structure in the harsh environment of impulse noise, processes the received signal through the infinite - norm weighted fractional lower - order moment, improves its noise suppression ability, enhances the convergence and direction - finding accuracy of the distributed array in the impulse - noise environment, breaks through the application limitations of the traditional distributed array direction - finding system where the performance deteriorates or even fails in the face of small - block beats and impulse - noise environments, and obtains a distributed array structure with a wider application range and better direction - finding performance.

[0032] 2. Compared with the traditional fox optimization algorithm that can only solve continuous optimization problems, the present invention introduces a quantum optimization mechanism to design a quantum fox optimization algorithm. By designing a quantum rotation - gate evolution strategy based on quantum coding, the update methods of the local optimal solution and the global optimal solution of the population are optimized. On the one hand, the search efficiency is improved, and on the other hand, the deficiency of the original search mechanism in the face of discrete optimization problems is made up, providing a new idea for solving integer programming problems. Description of the Drawings

[0033] Figure 1Schematic diagram of the array structure optimization method based on the quantum fox optimization algorithm;

[0034] Figure 2 Schematic diagram of the distributed array layout;

[0035] Figure 3 Relationship curve between the number of snapshots and the estimated root mean square error under different array structures;

[0036] Figure 4 Relationship curve between the number of snapshots and the estimated success probability under different array structures. Detailed implementation manners

[0037] The present invention will be further described below with reference to the accompanying drawings.

[0038] A distributed array structure optimization method based on the quantum fox optimization mechanism applicable to small sampling snapshot numbers and impulsive noise environments proposed by the present invention imitates the hunting of foxes and designs two stages of a development mechanism and an exploration mechanism, and then is further improved through quantum optimization theory. Introducing quantum coding and quantum evolution is beneficial to jumping out of local optimal solutions and improving convergence. At the same time, the present invention simulates the performance of actual direction finding, processes the received signals through the infinite norm weighted fractional lower order moments, improves its noise suppression ability, and uses the root mean square error between the source estimate value and the actual value as the optimization target to ensure the generation of a more adaptable distributed array structure under different environments, thereby further ensuring the effectiveness and superiority of the optimized sub-array layout.

[0039] Example 1, the technical solution is based on Figure 1 as shown, and specifically includes the following steps:

[0040] Step 1: Model the impulsive noise and source information in space. Set the noise in space as impulsive noise obeying the α-stable distribution, with its characteristic exponent α and symmetry parameter β. Set K known narrowband sources with wavelength λ in space, where the azimuth angles θ = [θ1, θ2,... θ K , -π ≤ θ k ≤ π. Suppose a distributed array includes M s sub-arrays, and each sub-array has M z array elements, then the total number of array elements of the entire array is M = M z × M s , the element spacing within the sub-array is d, the arrangement of the sub-arrays in space is one-dimensional, and the spacing between each sub-array is an integer multiple of the element spacing.

[0041] Step 2: Use the snapshot data received by the array to construct an infinite norm weighted fractional lower order moment matrix for direction estimation, and then obtain the root mean square error equation of direction estimation. Let the q-th snapshot data received be y(q) = A(θ)s(q) + n(q), where q is the snapshot number label, and q ∈ [1, Q], and Q is the maximum number of snapshots. In the formula, A(θ) is the steering vector matrix, s(q) is the signal vector, and n(q) is an independent and identically distributed impulse noise vector. Perform infinite norm normalization processing on the snapshot data where the infinite norm weighted fractional lower order moment matrix of the q-th snapshot data where C(q) is the fractional lower order moment matrix of the q-th snapshot data, C(q) = [c1(q), c2(q), …, c M (q)] is an M×M dimensional matrix, where c n (q) = [c 1n (q), c 2n (q), …, c Mn (q)] T , n = 1, 2, …, M, and its row and column element is where p0 is the lower order moment parameter, (·) * denotes taking the conjugate. Take the processed infinite norm weighted fractional lower order moment matrix as the covariance matrix, and then use the MUSIC algorithm for DOA estimation. After the covariance matrix R is eigen-decomposed, it can be divided into two subspaces: the signal subspace U S and the noise subspace U N , The signal subspace U S is composed of the eigenvectors corresponding to the larger eigenvalues, and these eigenvalues form a large eigenvalue diagonal matrix Σ S ; while the noise subspace U N is composed of the remaining eigenvectors, which correspond to the smaller eigenvalues, and these eigenvalues form another small eigenvalue diagonal matrix Σ N . When the noise subspace U N is right-multiplied by the covariance matrix R, we get Since the covariance matrix R S is full-rank and invertible, through derivation, it can be obtained that the steering vector A(θ) and the noise subspace U N are orthogonal, that is, A H (θ)U N = 0. Based on this orthogonality, the spatial spectrum function can be further derived Then, through spectrum peak search, the estimated azimuth value is obtained in the th experiment. Substitute the direction finding result into the objective function Among them is the solution to be obtained for the array position structure, K is the number of known signal sources in space, and θ k is the accurate azimuth value of the k-th signal source, is the azimuth estimation value of the k-th signal source at the th experiment when the array position is z.

[0042] Step 3: Initialize the quantum fox swarm to obtain the local optimal measurement position and the global optimal measurement position. First, assume that there are N quantum foxes in space, and each quantum fox is evenly distributed. Then, the quantum position of the n-th quantum fox in the t-th generation is The measurement equation is The measurement position is obtained The corresponding encoded array position is Among them, is a random number uniformly distributed between [0, 1], O is the maximum dimension of the solution space, and the maximum number of iterations is set to T. Map the array position corresponding to the position of the n-th fox obtained into the fitness function for evaluation, The fitness function of is Calculate the fitness value of the position of the n-th fox, where θ k is the accurate azimuth value of the k-th signal source, is the th experiment in the t-th generation, and the estimated direction of the k-th incoming wave obtained after the n-th fox executes the direction finding algorithm at the array position Then, record the optimal measurement position of the n-th fox up to the t-th generation as the local optimal position Record the measurement position with the best fitness up to the t-th generation as the global optimal position

[0043] Step 4: Update the quantum rotation angle and quantum position of each fox using the exploitation strategy and obtain the measurement position. For the n-th fox, set the hunting opportunity probability as a random number uniformly distributed between [0, 1]. If is greater than γ1, the fox moves towards the global optimal solution and calculates the sound propagation distance where the sound propagation time is defined as a random number distributed between [0, 1]. Find the sound speed through the best position Then the distance between the fox and the prey is half of the distance found through the sound wave After finding the distance between the fox and the prey, find the new position through the jump height. The average time of the n-th fox in the t-th generation is is a random number uniformly distributed between [0, 1], and the jump height is Define the o - th dimensional quantum rotation angle of the n - th quantum fox in the (t + 1)-th generation as If is less than γ1, then the o - th dimensional quantum rotation angle of the n - th quantum fox in the (t + 1)-th generation is defined as where is a random number uniformly distributed between [0, γ1], is a random number uniformly distributed between [γ1, 1]. The evolution method of the o - th dimensional quantum position of the n - th quantum fox is where is a random number uniformly distributed between [0, 1], b1 represents the probability of performing an inversion operation on this qubit when the quantum rotation angle is 0, and its value is a constant between [0, 1 / O]. Finally, the corresponding position is obtained by measuring the quantum position Its measurement equation is where is a random number uniformly distributed between [0, 1], n = 1, 2, …, N, o = 1, 2, …, O. Substitute the measured position of the n - th fox into the fitness function to obtain the fitness value If then the local optimal measured position of the n - th fox up to the (t + 1)-th generation is Otherwise, Select the optimal position from the N local optimal measured positions up to the (t + 1)-th generation and set it as the global optimal measured position up to the (t + 1)-th generation

[0044] Step 5: Update the quantum rotation angle and quantum position of each fox using an exploration strategy and obtain the measured position. During the exploration phase, the foxes hunt using random walks. To ensure that the foxes move towards the optimal position, use a small variable and variable a to control the search, is the minimum average time of the n - th individual up to the t - th generation, is a random number uniformly distributed between [0, 1], is the average time of the n - th individual in the t - th generation. The random variable where T is the maximum number of iterations. Define the o - th dimensional quantum rotation angle of the n - th quantum fox in the (t + 1)-th generation as where is a random number uniformly distributed between [0, 1], and then define the evolution method of the o - th dimensional quantum position of the n - th quantum fox as where is a random number uniformly distributed between [0, 1]. b1 represents the probability of performing an inversion operation on this qubit when the quantum rotation angle is 0, and its value is a constant between [0, 1 / O]. Finally, the corresponding position is obtained by measuring the quantum position, and its measurement equation is where is a random number uniformly distributed between [0, 1], n = 1, 2, …, N, o = 1, 2, …, O.

[0045] Step 6: Substitute the measurement position obtained by the nth fox into the fitness function to obtain the fitness value If then determine the local optimal measurement position until the (t + 1)th generation Otherwise the value remains unchanged. Select the optimal position from the N local optimal measurement positions until the (t + 1)th generation and set it as the global optimal measurement position until the (t + 1)th generation

[0046] Step 7: Determine whether the maximum number of iterations is reached. If so, output the global optimal measurement position and convert it into array layout structure information; otherwise, let t = t + 1 and return to Step 4 to continue the iteration.

[0047] Example 2: Simulate the distributed array layout optimization design method based on the quantum fox optimization mechanism, and its parameter design is as follows: The total number of array elements M = 80, the number of array elements in each sub - array M z = 8, the number of sub - arrays M s = 10, the element spacing in the sub - array The schematic diagram of the array structure optimization method based on the quantum fox optimization algorithm is as Figure 1 , and the overall arrangement of the distributed array is as Figure 2 . The number of quantum foxes N = 200, the spatial dimension O = 72, the maximum number of iterations T = 200, the selection probability γ1 = 0.18, the gravitational acceleration γ g= 9.81. Then, a reference signal source and the corresponding direction-of-arrival (DOA) estimation method for evaluating the quality of the generated distributed array are set. The total number of reference signal sources is K = 4, and the incident directions are θ1 = 58.3°, θ2 = 36.5°, θ3 = 10.1°, and θ4 = -25.2° respectively. The spatial noise is additive impulsive noise, the characteristic exponent α = 1.5, the symmetry parameter β = 0, and the generalized signal-to-noise ratio is 5 dB. For the DOA estimation method of the signal source, the MUSIC algorithm based on the infinite norm weighted fractional lower-order moments is adopted. The step size of the algorithm spectrum peak search is set to 0.1°, and the number of snapshots of the received data is set to 10. Finally, the optimal value is obtained by taking the best value through multiple tests, and the layout structure of the final distributed array system is obtained. The spacing between adjacent sub-arrays is D = [36d, 77d, 139d, 70d, 56d, 133d, 97d, 176d, 195d].

[0048] To verify the superiority of the obtained distributed array layout in the DOA measurement performance in the actual environment, the distributed array layout obtained based on the quantum fox optimization mechanism is compared with the uniform linear array and the distributed array structures with sub-array spacings of 15d and 30d in terms of the DOA measurement performance in the environments of small snapshots, small generalized signal-to-noise ratio, and impulsive noise. The spatial source and noise environment are both set as above.

[0049] Figure 3 It shows the influence of different array structures on the DOA measurement performance under different sampling snapshot numbers when the generalized signal-to-noise ratio is 5 dB under impulsive noise, with the root mean square error as the measurement index of the DOA accuracy. 200 independent Monte Carlo experiments are carried out at each snapshot number. It can be seen from this that in the case of low generalized signal-to-noise ratio, as the number of snapshots increases, the DOA measurement performance gradually improves. However, when the number of snapshots is small, the array structure generated based on the quantum fox optimization mechanism has obvious superiority and higher DOA accuracy.

[0050] Figure 4 It shows the influence of DOA measurement with different snapshot numbers on the estimation success probability when the generalized signal-to-noise ratio is 5 dB under impulsive noise. It is stipulated that when the error between the angle estimation value and the actual target value is less than 1.5°, it is determined as an estimation success. 200 independent Monte Carlo experiments are carried out at each snapshot number. It can be seen from this that the array generated based on the quantum fox optimization mechanism has stronger noise suppression ability and higher DOA accuracy. When the number of snapshots is large, most array structures can reach a relatively ideal estimation probability. However, when the number of snapshots is small, only the array structure generated by the proposed method can maintain a high estimation probability, and the estimation results of the other array forms deteriorate or even fail to varying degrees, indicating that the proposed array layout optimization method has stronger adaptability and breaks through the limitations of the existing array layout structures.

[0051] Embodiment 3: The functions of an optimization system for the distributed array structure of a quantum fox mechanism according to the present invention can be illustrated by the foregoing optimization method for the distributed array structure of a quantum fox mechanism. Therefore, for the parts not detailed in the system embodiment, reference can be made to the above method embodiment, which will not be elaborated herein.

[0052] Embodiment 4: An electronic device according to the present invention includes a memory and a processor. Among them, the memory is used to store a program; the processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the foregoing optimization method for the distributed array structure of a quantum fox mechanism.

[0053] Embodiment 5: A computer-readable storage medium according to the present invention stores a computer program thereon. When the computer program is executed by a processor, it implements the foregoing optimization method for the distributed array structure of a quantum fox mechanism.

[0054] It should be noted that the logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in combination with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0055] It should be understood that each part of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0056] Embodiment 6: A computer program product of the present invention, when the computer program product runs on a computer, executes an optimization method for a distributed array structure of a quantum fox mechanism. Wherein, the computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable devices.

[0057] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A distributed array structure optimization method for a quantum fox mechanism, characterized in that: The following steps are involved: Step 1: Model the impact noise and source information in space; Step 2: Use the snapshot data received by the array to construct an infinite norm weighted fractional low-order moment matrix for direction estimation, and obtain the root mean square error equation of the direction estimation; Step 3: Initialize the quantum fox group to obtain the local optimal measurement position and the global optimal measurement position; Step 4: Use the development strategy to update the quantum rotation angle and quantum position of each fox and obtain the measured position; Step 5: Use the exploration strategy to update the quantum rotation angle and quantum position of each fox and obtain the measured position; Step 6: Substitute the measurement position obtained by the nth fox of the t+1th generation into the fitness function to obtain the fitness value, and judge the fitness value to obtain the local optimal measurement position and the global optimal measurement position; Step 7: Determine whether the maximum number of iterations has been reached. If so, output the global optimal measurement position and convert it into array layout structure information; otherwise, set t=t+1 and return to step 4 to continue iterating.

2. The distributed array structure optimization method of a quantum fox mechanism according to claim 1, characterized in that: The step 1 is specifically as follows: The noise in the space is set to be an impulse noise that obeys the α-stable distribution, with a characteristic index α and a symmetry parameter β; K known narrowband signal sources with a wavelength of λ are set in the space, where the azimuth angle θ=[θ1,θ2,...θ K ],-π≤θ k ≤π; suppose a distributed array contains M s sub-arrays, each with M z The total number of elements in the entire array is M = M z ×M s , the spacing between array elements in a subarray is d, the subarray is arranged in space in a one-dimensional manner, and the spacing between each subarray is an integer multiple of the spacing between array elements.

3. The distributed array structure optimization method of a quantum fox mechanism according to claim 1, characterized in that: The step 2 is specifically as follows: Assume that the received qth snapshot data is y(q)=A(θ)s(q)+n(q), where q is the snapshot number, and q∈[1,Q], Q is the maximum snapshot number, A(θ) is the steering vector matrix, s(q) is the signal vector, and n(q) is the independent and identically distributed impulse noise vector; perform infinite norm normalization on the snapshot data in, Infinity norm weighted fractional low-order moment matrix of the qth snapshot data Where C(q) is the fractional low-order moment matrix of the qth snapshot data, C(q) = [c1(q), c2(q), ..., c M (q)] is an M×M dimensional matrix, where c n (q) = [c 1n (q),c 2n (q),...,c Mn (q)] T , n=1,2,…,M, its Line The elements of the column are Where p0 is the low-order moment parameter, (·) * Indicates taking conjugate; the processed infinite norm weighted fractional low-order moment matrix is ​​used as the covariance matrix, and then the MUSIC algorithm is used for DOA estimation. After eigendecomposition, the covariance matrix R is divided into two subspaces: signal subspace U S and the noise subspace U N , Signal subspace U S It consists of eigenvectors corresponding to larger eigenvalues, which form a large eigenvalue diagonal matrix Σ S ; and the noise subspace U N The remaining eigenvectors correspond to smaller eigenvalues, which form another small eigenvalue diagonal matrix ∑ N ; When the noise subspace U N After right multiplication with the covariance matrix R, we get Since the covariance matrix R S Full rank and reversible, after derivation, the steering vector A(θ) and the noise subspace U N is orthogonal, that is, A H (θ)U N =0; Based on this orthogonality, the spatial spectrum function is further derived Then search the peaks The estimated position value was obtained by experiment Substitute the direction finding results into the objective function in, is the array position structure to be solved, K is the number of known sources in space, θ k is the accurate position value of the kth source, When the array position is z The estimated value of the direction of the kth source in this experiment.

4. The distributed array structure optimization method of a quantum fox mechanism according to claim 1, characterized in that: The step three is specifically as follows: First, assume that there are N quantum foxes in space, and each quantum fox is evenly distributed. Then the quantum position of the nth quantum fox in the tth generation is n=1,2,…,N,o=1,2,…,O,the measurement equation is Get the measurement position The corresponding encoded array position is in, is a random number that follows a uniform distribution between [0,1]. O is the maximum dimension of the solution space, and the maximum number of iterations is set to T; the array position mapped to the position of the nth fox Bring in the fitness function for evaluation, The fitness function is Calculate the position of the nth fox The fitness value of k is the accurate position value of the kth source, For the tth generation The nth fox in the array is at position The estimated direction of the kth incoming wave is obtained after executing the direction finding algorithm, and then the optimal measurement position of the nth fox up to the tth generation is recorded as the local optimal position The measurement position with the best fitness up to the tth generation is recorded as the global optimal position 5. The distributed array structure optimization method of a quantum fox mechanism according to claim 1, characterized in that: The step 4 is specifically as follows: For the nth fox, set the probability of hunting opportunity is a random number between [0,1] that follows a uniform distribution. If If it is greater than γ1, the fox moves towards the global optimal solution and calculates the sound propagation distance The sound propagation time Defined as a random number distributed between [0,1], the speed of sound is found through the optimal position The distance between the fox and the prey is half the distance found by the sound waves. After finding the distance between the fox and the prey, the fox finds the new position by jumping height. The average time for the nth fox in the tth generation is is a random number uniformly distributed between [0,1], and the jump height is The o-th dimension quantum rotation angle of the n-th quantum fox of the t+1th generation is defined as like is less than γ1, then the o-th dimension quantum rotation angle of the n-th quantum fox of the t+1th generation is defined as in, is a random number that follows a uniform distribution between [0,γ1], is a random number that follows a uniform distribution between [γ1,1]; the evolution of the o-th dimension quantum position of the n-th quantum fox is in is a random number that follows a uniform distribution between [0,1], b1 represents the probability of performing a reversal operation on the quantum bit when the quantum rotation angle is 0, and its value is a constant between [0,1 / O]; finally, the corresponding position is obtained by measuring the quantum position The measurement equation is in is a random number uniformly distributed between [0,1], n=1,2,…,N, o=1,2,…,O; the measured position of the nth fox Substitute into the fitness function to get the fitness value like Then the local optimal measurement position of the nth fox until generation t+1 is otherwise, The best position is selected from the N local best measurement positions up to generation t+1 and is set as the global best measurement position up to generation t+1 6. The distributed array structure optimization method of a quantum fox mechanism according to claim 1, characterized in that: The step five is specifically as follows: In the exploration phase, the fox uses random walks to hunt. In order to ensure that the fox moves to the optimal position, a small variable is used and variable a to control the search, is the minimum average time of the nth individual since the tth generation, is a random number that follows a uniform distribution between [0,1]. is the average time of the nth individual in the tth generation; random variable Where T is the maximum number of iterations; the o-th dimension quantum rotation angle of the n-th quantum fox in the t+1th generation is defined as in is a random number that follows a uniform distribution between [0,1], and then the evolution of the o-th dimension quantum position of the n-th quantum fox is defined as in is a random number that follows a uniform distribution between [0,1], b1 represents the probability of performing a reversal operation on the quantum bit when the quantum rotation angle is 0, and its value is a constant between [0,1 / O]; finally, the corresponding position is obtained by measuring the quantum position, and its measurement equation is in is a random number from [0,1] that follows a uniform distribution, n = 1, 2, …, N, o = 1, 2, …, O.

7. The distributed array structure optimization method of a quantum fox mechanism according to claim 1, characterized in that: The step six is ​​specifically as follows: The measured position of the nth fox in the t+1 generation is Substitute the fitness function to get the fitness value like Then determine the local optimal measurement position until generation t+1 otherwise The value remains unchanged; the optimal position is selected from the N local optimal measurement positions until the t+1 generation and set to the global optimal measurement position until the t+1 generation 8. A computer device / equipment / system comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images