A small snapshot wave direction estimation method, system and storage medium of a quantum dream optimization mechanism under impact noise

Abstract

The application belongs to the field of array signal processing, and discloses a small snapshot direction of arrival estimation method and system of a quantum dream optimization mechanism under impulse noise and a storage medium. First, an array receiving model under impulse noise environment is established to obtain a receiving data matrix; a robust covariance matrix is calculated based on the data matrix, and a hybrid kernel correlation entropy covariance matrix is constructed by using at least two different kernel functions, and an enhanced correlation entropy covariance matrix is obtained by fusing the two; then, eigenvalue decomposition is performed on the enhanced correlation entropy covariance matrix to establish a weighted signal subspace fitting target function; a quantum dream optimization algorithm is used for solving, and the iteration process is divided into an exploration stage and a development stage by setting a stage switching threshold: only local optimal information of a subgroup is used for position updating in the early stage, and only global optimal information is used for position updating in the late stage; finally, an angle parameter corresponding to a globally optimal position is output. The application can effectively estimate the angle of a coherent source signal under the conditions of impulse noise and small snapshots, has global convergence and real-time performance, and significantly improves the robustness and estimation accuracy in a complex electromagnetic environment.

Description

Technical Field

[0001] This invention belongs to the field of array signal processing, specifically relating to a small snapshot direction-of-arrival estimation method, system, and storage medium for a quantum dream optimization mechanism under impulse noise. Background Technology

[0002] Direction of Arrival (DOA) estimation is one of the core problems in array signal processing, widely used in target localization, beamforming, and spatial spectrum estimation. Traditional DOA estimation methods, such as MUSIC and ESPRIT, perform well in Gaussian noise environments, but in practical engineering, non-Gaussian impulse noise, such as electromagnetic interference, sea clutter, and transient discharges, is frequently encountered. The distribution of signal and noise often exhibits significant non-Gaussian characteristics, specifically, its probability density function has sharper peaks and thicker tails than Gaussian distributions, leading to a severe deterioration in the performance of traditional methods. Since the core of these methods relies on eigenvalue decomposition of the covariance matrix, their estimation accuracy is largely limited by the number of available snapshots. Only with a sufficient number of snapshots can they exhibit ideal performance, resulting in low real-time performance and high computational cost. To reduce the computational cost of DOA estimation and improve the real-time performance of the system, small-snapshot DOA estimation has received widespread attention. However, small-snapshot DOA estimation often requires a high signal-to-noise ratio, and the eigenvalue of the impulse noise should not be too large, otherwise its robustness is poor. Therefore, it is essential to design a wave arrival direction estimation method with high accuracy and applicable to impulsive noise environments.

[0003] A review of existing literature reveals that early related research largely revolved around Gaussian noise models, and the improvement ideas also stemmed from this. For example, Liu, T.-H. et al. systematically proposed a subspace algorithm based on fractional lower order statistics in "A subspace-based direction finding algorithm using fractional lower order statistics". It directly defined the covariance matrix based on fractional lower order moments and proved that the matrix still contains subspace information of the signal under impulse noise, thus enabling the MUSIC and ESPRIT algorithms to be extended to impulse noise environments. Shan Zebiao et al. published "DOA estimation of coherent signals under impulse noise background" in Electronic Measurement Technology (2020, 42(11):2600-2606), proposing a FLOM-MMUSIC method. This method can suppress the influence of impulse noise on array data by obtaining the fractional lower order matrix based on the array received data. By rearranging the matrix conjugates, it solves the problem of loss of array covariance matrix rank caused by the presence of coherent signals, and has good estimation performance under low signal-to-noise ratio and small snapshot conditions. In their paper "A Uniform Linear Array DOA Estimation Method Based on Unitary Transformation" published in the *Journal of Information Engineering University* (2022, 23(01):24-27), Gao Weigang et al. proposed the UNITARY-MUSIC method. By performing a unitary transformation on the covariance matrix of the received signal, the complex domain operations are transformed into real domain operations, effectively reducing complexity and improving resolution performance. However, the number of snapshots required by the above two methods is around 100 to 200, which is still relatively large. In their paper "A Fast Estimation Method for Stochastic Maximum Likelihood DOA Based on Membrane Computation Optimization" published in *High Technology Communications* (2019, 29(09):833-840), Xiang Changbo et al. proposed a membrane computation method based on particle swarm optimization and artificial bee colony optimization. This method has the ability to perform global and local searches simultaneously, and its estimation accuracy is far superior to that of the traditional spatial spectrum algorithm, improving the convergence speed. However, it is not applicable to impulse noise. Summary of the Invention

[0004] The purpose of this invention is to provide a method, system, and storage medium for estimating the direction of arrival (DOA) of a quantum dream optimization mechanism under impulsive noise conditions, thereby solving the problems of insufficient accuracy and easy entrapment into local optima in existing ODA estimation methods under impulsive noise environments.

[0005] The objective of this invention is achieved through the following technical solution:

[0006] A method for estimating direction-of-arrival (DOA) of a quantum dream optimization mechanism under impulsive noise, comprising the following steps:

[0007] Step 1: Establish an array signal receiving model under impulse noise environment and obtain the array received data matrix;

[0008] Step 2: Calculate the robust covariance matrix based on the array received data matrix; construct a hybrid kernel correlation entropy covariance matrix using at least two different kernel functions; fuse the robust covariance matrix with the hybrid kernel correlation entropy covariance matrix to obtain the enhanced correlation entropy covariance matrix;

[0009] Step 3: Perform eigenvalue decomposition on the enhanced robust covariance matrix and establish a small snapshot weighted signal subspace fitting equation based on the enhanced hybrid kernel correlation entropy. The objective of the equation is to solve for the angle parameters that minimize the weighted signal subspace fitting criterion.

[0010] Step 4: The quantum dream optimization algorithm is used to solve the weighted subspace fitting objective function. The individuals in the quantum dream optimization algorithm population are updated in the search space to obtain the optimal solution.

[0011] Step 5: Set a phase switching threshold to divide the iteration process into an exploration phase and a development phase: When the current iteration count is less than or equal to the phase switching threshold, only the exploration phase is executed, and the position of individuals in the exploration phase is updated based on the local optimum information of their respective subgroups; when the current iteration count is greater than the phase switching threshold but less than the maximum iteration count, only the development phase is executed, and the position of individuals in the development phase is updated based on the global optimum information.

[0012] Step 6: Determine if the current iteration count has reached the maximum iteration count. If it has, output the angle parameter corresponding to the current global optimal position as the direction of arrival estimation result; otherwise, update the iteration count and return to step 4.

[0013] Furthermore, the array signal receiving model described in step 1 is based on a uniform linear array, which contains M array elements with an element spacing of d, and receives incident signals from N far-field narrowband signal sources, with each signal source having an incident angle of d. The array signal receiving model received the first... The narrowband signal of the second snapshot is The received data matrix is :

[0014]

[0015] in, For the guiding matrix, For the first Sub-sampling signal vector, For the first Sub-sampling impact noise vector This indicates that the signal received by the signal receiving model is the first... The next snapshot is a narrowband signal, where K represents the total number of snapshots. This represents a snapshot narrowband signal vector received by the signal receiving model. This represents the vector of a single snapshot signal. This represents the impact noise vector of a single snapshot.

[0016] Furthermore, the enhanced correlation entropy covariance matrix described in step 2 for:

[0017]

[0018] in, These are the weighted combination coefficients. ; Let the robust covariance matrix be denoted by the th matrix. Line 1 Column elements are ,in When it satisfies Hermitian symmetry, that is , represent The Middle All elements in the row, represent The Middle All elements in the row, This is an element-wise multiplication operation. To perform conjugate operations, For the median operation, , ; Let the relevant entropy covariance matrix be represented, and its i-th Line 1 The elements of the column are ,in , , , These are the weighting coefficients. , For matrix The first in Line 1 Column elements, For matrix The first in Line 1 Column elements, and is the kernel function for the correlation entropy.

[0019] Furthermore, in step 3, eigenvalue decomposition is performed on the enhanced robust covariance matrix to establish an angle estimation equation based on the small snapshot weighted signal subspace fitting equation of the enhanced hybrid kernel correlation entropy.

[0020]

[0021] in, The estimated vector of the incoming wave angle that maximizes the projection trace of the weighted signal subspace; It is a diagonal weighted matrix, whose diagonal elements are determined by eigenvalues; To find the trace function; For guiding matrix The mapping matrix; The signal subspace matrix; is the conjugate transpose of the signal subspace matrix; This is the conjugate transpose operator.

[0022] Furthermore, in step 4 of the quantum dream optimization algorithm, the first... The second iteration The quantum position of each individual is ,in, , D is the dimension of the search space, an integer. Let be the number of iterations; the th The second iteration The mapping position corresponding to the quantum position of each individual is: The mapping rule is ,in, , Mapping position of quantum individual The upper limit of dimensional variables, Mapping position of quantum individual Lower bound of the dimensional variable; calculate the dimensional variable. The fitness value of all positions in the quantum population is , among which, the The generation The expression for the fitness value of the mapped position of a quantum individual is: The population is grouped according to the mean of quantum positions, and each subpopulation evolves independently; the quantum position of the quantum individual with the highest fitness within a group is denoted as . As the collective memory of the group; record the iteration to the . The global optimal fitness value up to date is The quantum position of the corresponding individual is the globally optimal quantum position. .

[0023] Furthermore, the stage switching threshold in step 5 is... When the current iteration number At that time, the exploration phase is executed; when During the development phase, the initial timing command is executed. , This represents the maximum number of iterations.

[0024] During the exploration phase: for each quantum individual within the group Random selection Each dimension is a "forgotten dimension," and only the location information of the "forgotten dimension" is updated; with adaptive probability. Implement the forgetting supplementation strategy: in the first In each dimension of forgetting, the quantum rotation angle is calculated based on the collective memory. Applying quantum rotation gates to the first The generation Group 1 The quantum position of an individual To supplement memories and obtain the first The generation A quantum individual in the dimension of forgetting The quantum position on ,in, , To explore the magnitude of the rotation angle during the exploration phase, To supplement the memory coefficients; with preset probabilities Implementing the Dream Sharing Strategy: In the first In each forgetting dimension, the quantum position of another quantum individual in the same subgroup is randomly selected to calculate the quantum rotation angle, and the update is performed in the same way as the forgetting replenishment strategy, while the position information in other dimensions remains unchanged.

[0025] Furthermore, during the development phase, all quantum individuals are randomly selected. The first dimension is the forgetting dimension; only the position information of the forgetting dimension is updated. Calculate the quantum rotation angle based on the globally optimal quantum position in each forgetting dimension. Applying quantum rotation gates to the first The generation The memory is supplemented by the quantum position of each individual to obtain the first... The generation The dimension of forgetting of individual quantum entities quantum position ,in, , To develop the magnitude of the rotation angle.

[0026] Furthermore, after the first iteration is completed, the [number]th iteration is obtained. The first generation Individuals The quantum position in the dimensional search space is Calculate the corresponding mapping position and its fitness value ; Calculate the first The maximum fitness value in a quantum population and compare it with the current global optimal fitness value. Compare; if Then update the global highest fitness value and record it. The corresponding quantum position is the globally optimal quantum position; otherwise, it is recorded. The corresponding quantum position is the globally optimal quantum position.

[0027] A computer system includes a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of a small snap-to-arrival direction estimation method for a quantum dream optimization mechanism under impulse noise.

[0028] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a small snap-shot direction-of-arrival estimation method for a quantum dream optimization mechanism under impulse noise.

[0029] The beneficial effects of this invention are as follows:

[0030] Compared with existing technologies, this invention addresses the prominent shortcomings of existing small snapshot DOA estimation methods, such as insufficient adaptability in complex electromagnetic environments and robustness degradation in low signal-to-noise ratio and impulsive noise scenarios. The DOA estimation method based on the quantum dream optimization mechanism proposed in this study on small snapshot DOA estimation under impulsive noise can ensure effective DOA estimation of coherent source signals in low signal-to-noise ratio and impulsive noise environments, while achieving global convergence and real-time performance. Attached Figure Description

[0031] Figure 1 A schematic diagram of the operation of the direction-of-arrival estimation method based on the quantum dream optimization mechanism under impact noise proposed in this invention.

[0032] Figure 2 Success probability diagram of DOA estimation of coherent signals under different generalized signal-to-noise ratios in a strong impact noise environment.

[0033] Figure 3 Root mean square error curves of DOA estimation for coherent signals under different generalized signal-to-noise ratios in a high-impact noise environment.

[0034] Figure 4 Probability of successful DOA estimation for coherent signals under different intensity impulse noise in an environment with the same signal-to-noise ratio.

[0035] Figure 5The root mean square error curves of DOA estimation for coherent signals under different intensity impulse noise in an environment with the same signal-to-noise ratio. Detailed Implementation

[0036] The present invention will now be further described with reference to the accompanying drawings.

[0037] This invention discloses a small-shot direction-of-arrival (DOA) estimation method, system, and storage medium under impulsive noise using a quantum dream optimization mechanism. To address model mismatch caused by impulsive noise and eliminate signal correlation, this invention introduces a hybrid kernel function strategy combining hyperbolic tangent and Huber functions. A convex combinatorial structure is used to construct an enhanced correlation entropy covariance matrix based on the hybrid kernel, yielding a small-shot weighted signal subspace fitting equation based on the enhanced hybrid kernel correlation entropy. To prevent getting trapped in local optima, this invention designs a quantum dream optimization method to efficiently solve the subspace fitting equation based on the enhanced hybrid kernel correlation entropy. The invented small-shot DOA estimation method effectively locates coherent source energy under impulsive noise. Figure 1 The specific steps are as follows:

[0038] Step 1: Establish a signal reception model under impact noise environment.

[0039] Consider a system containing impact noise in an environment A uniform linear array with n elements, the element spacing being 1. Suppose there is Several far-field narrowband signal sources, with incident angles of respectively The frequency is The number of quick shots is The incident signal and noise are uncorrelated; taking the first array element as the reference array element, then the... Individual elements The signal received at time t can be represented as , ,in, Indicates in The incident direction at any moment is Narrowband signal, For complex units, Indicates in Time of the first The impulse noise on the nth array element, then the array receives the nth... The narrowband signal of the second snapshot is The received data matrix is In the formula Let be the guiding matrix, where the first... The guide vectors are , For the first Sub-sampling signal vector, For the first Subsequent snapshot sampling of the impact noise vector.

[0040] Step 2: Construct an enhanced correlation entropy covariance matrix based on a hybrid kernel.

[0041] Computation array receive data matrix The robust covariance matrix is ​​obtained by taking the median of the element-wise product. , its first Line 1 The column elements are calculated as follows: ,in When it satisfies Hermitian symmetry, that is , represent The Middle All elements in the row, represent The Middle All elements in the row, This is an element-wise multiplication operation. To perform conjugate operations, For the median operation, , Then, construct the correlation entropy covariance matrix based on the hybrid kernel. , of which Line 1 Column elements It can be specifically expressed as ,in , , , These are the weighting coefficients. , For matrix The first in Line 1 Column elements, For matrix The first in Line 1 Column elements, and The kernel function of the correlation entropy. , For scale parameters, , This is the scaling parameter. To improve performance and ensure the stability of the mechanism, the relevant entropy covariance matrix is... and By performing a convex combination, we obtain the enhanced correlation entropy covariance matrix. , These are the weighted combination coefficients. .

[0042] Step 3: Establish a small snapshot weighted signal subspace fitting equation based on enhanced hybrid kernel correlation entropy.

[0043] For the enhanced correlation entropy covariance matrix By performing eigenvalue decomposition, we can obtain , where the signal subspace diagonal matrix Depend on The noise subspace is composed of several large signal eigenvalues. The remaining The eigenvalues ​​form a diagonal matrix. The angle estimation equation based on the small snapshot weighted signal subspace fitting equation of the enhanced hybrid kernel correlation entropy is: ,in, It is a diagonal weighted matrix, and its first... The diagonal elements are , , It is the first Each signal characteristic value, It is the largest signal characteristic value. It is an inverse intensity parameter. As a constant enhancement factor, To find the trace function, For guiding matrix The mapping matrix, where This represents finding the inverse of a matrix.

[0044] Step 4: Initialize the quantum dream optimization algorithm population search mechanism, calculate the fitness of individuals, group the population, and determine the optimal quantum position and the global optimal position of each subgroup.

[0045] For a The optimization problem of dimension , where the number of individuals in the quantum population is . Randomly initialize the quantum population, with a maximum number of iterations. The number of iterations is marked as an integer. Initial command The population is divided according to the mean of quantum positions. There are 1 group, and each subgroup evolves independently. The second iteration The quantum position of each individual is ,in , , The corresponding mapping position is The specific mapping rules are as follows: ,in, , Mapping position of quantum individual The upper limit of dimensional variables, Mapping position of quantum individual The lower bound of the dimension variable. Calculate the _th_... The fitness value of all positions in the quantum population is , of which The generation The expression for the fitness value of the mapped position of a quantum individual is: Let the quantum position of the quantum individual with the highest fitness within the group be denoted as . This serves as the group's "collective memory." The iteration is recorded up to the [number missing]. The highest global fitness value up to date is The quantum position of the corresponding individual is the globally optimal quantum position. .

[0046] Step 5: Exploring the optimization mechanism of quantum dreams.

[0047] Set a number of iterations to divide the different stages. ,when At this point, the exploration phase of the quantum dream optimization mechanism is conducted. This exploration phase simulates memory-based exploration behavior in dreams, simulating the dream process once per iteration, and gradually approaching the optimal solution through repeated iterations. This involves each quantum individual within the group. Random selection Each dimension is a "forgotten dimension," and only the positional information of the "forgotten dimension" is updated, while the positional information of other dimensions remains unchanged. This is achieved using adaptive probability. Implement the forgetting supplementation strategy: in the first On the dimension of forgetting, the quantum rotation angle is calculated based on collective memory. Applying quantum rotation gates to the first The generation Group 1 The quantum position of an individual To supplement memories and obtain the first The generation A quantum individual in the dimension of forgetting Quantum position on: In the formula , , , The magnitude of the rotation angle. Supplementary coefficients for memory. Based on probability. Implement the dream-sharing strategy: mimicking the process of humans sharing and exchanging ideas in dreams, allowing quantum individuals to learn from other random quantum individuals in the same group in the forgetting dimension, with the update method being the same as above, while the position information in other dimensions remains unchanged.

[0048] After the exploration phase is completed, the first [item] will be obtained. The first generation Individuals The quantum position in the dimensional search space is Calculate the corresponding mapping position and its fitness value Find the first generational quantum population fitness The maximum value in ,and Comparison, if Then update the global highest fitness value and record it. The corresponding quantum position is the globally optimal quantum position. Conversely, record The corresponding quantum position is the globally optimal quantum position. .

[0049] Step Six: Development of the Quantum Dream Optimization Mechanism.

[0050] when At this time, the development phase of the quantum dream optimization mechanism is underway. During this development phase, grouping is no longer performed; all quantum individuals are randomly selected. The first dimension is the forgetting dimension; only the positional information of the forgetting dimension is updated, while the positional information of other dimensions remains unchanged. In the first... Calculate the quantum rotation angle based on the globally optimal quantum position in each forgetting dimension. Applying quantum rotation gates to the first The generation The memory is supplemented by the quantum position of each individual to obtain the first... The generation A quantum individual in the dimension of forgetting quantum position In the formula , , The magnitude of the rotation angle. Supplementary coefficients for memory.

[0051] After the development phase is completed, the first The first generation Individuals The quantum position in the dimensional search space is Calculate the corresponding mapping position and its fitness value Find the first generational quantum population fitness The maximum value in ,and Comparison, if Then update the global highest fitness value and record it. The corresponding quantum position is the globally optimal quantum position. Conversely, record The corresponding quantum position is the globally optimal quantum position. .

[0052] Steps five and six specifically involve setting the total number of iterations to 80 during the algorithm iteration process. The first 64 iterations execute only the exploration phase, skipping the development phase, and expanding the search space through forgetting and supplementation strategies and dream-sharing strategies. The last 16 iterations execute only the development phase, skipping the exploration phase, and performing a refined search through a global memory strategy. This phased mechanism of "exploration first, development later" effectively balances global search and local mining capabilities, significantly improving DOA estimation performance under conditions of small snapshots and impact noise.

[0053] Step 7: Determine if evolution terminates.

[0054] Determine if the current iteration count has reached the maximum iteration count. If the target is not reached, increment the value by 1 and return to step four; if the target is reached, terminate the iteration and output the mapping position corresponding to the global optimal quantum position in the last generation as the direction of arrival estimation result.

[0055] Example:

[0056] The quantum dream-based optimization method designed in this invention is abbreviated as "QDOA-HE-WSSF". In the simulation experiment, two multipath coherent signals are respectively generated by... Incident from two directions at a distance of On a uniform linear array, where The wavelength is denoted as . The simulation parameters are designed as follows: number of antenna array elements. =20, Number of sources =2, quantum population size Maximum number of iterations Quick shot number Dimensions Each experiment is run 100 times, with weighting coefficients... =0.5, scale parameter , Weighted combination coefficient Inverse intensity parameter constant enhancement factor Exploration phase Number of groups Adaptive probability The rotation angle magnitude of the forgetting supplementation strategy The rotation angle amplitude of the dream-sharing strategy Rotation angle amplitude during development phase Memory compensation coefficient in the two stages .

[0057] The comparison algorithms used in the simulation experiment specifically refer to the FLOM-MMUSIC method proposed by Shan Zebiao et al. in "Estimation of DOA of Coherent Signal under Impact Noise Background" published in "Electronic Measurement Technology", and the UNITARY-MUSIC method proposed by Gao Weigang et al. in "A Method for Estimating DOA of Uniform Linear Array Based on Unitary Transformation" published in "Journal of Information Engineering University".

[0058] Figure 2 The characteristic index of the medium-impact noise is 1.0, and the generalized signal-to-noise ratio varies from -5dB to 15dB. The estimation is considered successful when the DOA estimation result differs from the incident direction by no more than 2°. Figure 2 As can be seen, in a strong impact noise environment, the success probability of QDOA-HE-WSSF is better than that of the comparative method under various high signal-to-noise ratios, and the success rate of QDOA-HE-WSSF is at a high level.

[0059] Figure 3 The characteristic index of the medium-impact noise is 1.0, and the generalized signal-to-noise ratio varies from -5dB to 15dB. From Figure 3 As can be seen, in a fixed environment with strong impact noise, the estimation curve of QDOA-HE-WSSF is at a low level, indicating that its estimation accuracy is much higher than that of the comparison method.

[0060] Figure 4 The generalized signal-to-noise ratio is 10dB, and the impulse noise characteristic index varies from 0.8 to 1.6. The estimation is considered successful when the DOA estimation result differs from the incident direction by no more than 2°. From... Figure 4 As can be seen, in a fixed low signal-to-noise ratio environment, the success probability of QDOA-HE-WSSF is better than that of the comparative method under various impulse noise conditions, and the success rate of QDOA-HE-WSSF is at a high level.

[0061] Figure 5 The generalized signal-to-noise ratio is 10dB, and the impulse noise characteristic index varies from 0.8 to 1.6. From Figure 5 As can be seen, in a fixed low signal-to-noise ratio environment, the estimation curve of QDOA-HE-WSSF is at a low level, indicating that its estimation accuracy is much higher than that of the comparison method.

[0062] In particular, in some preferred embodiments of the present invention, a computer device is also provided, including a memory and a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the small snap wave direction estimation method of the quantum dream optimization mechanism under impulse noise described in any of the above embodiments.

[0063] In some other preferred embodiments of the present invention, a computer-readable storage medium is also provided, on which a computer program / instruction is stored, wherein when the computer program is executed by a processor, the steps of the small snap wave direction estimation method of the quantum dream optimization mechanism under impulse noise described in any of the above embodiments are implemented.

[0064] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the small snap wave direction estimation method embodiment of the quantum dream optimization mechanism under impulse noise described above, which will not be repeated here.

[0065] Computer-readable storage media encompass a variety of types, including persistent and non-persistent, portable and fixed. These media store information using different technologies, and the content can be machine instructions, data structures, program modules, or other types of data. Some typical examples of computer storage media include: phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), various types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory and other storage technologies, optical storage media such as CD-ROM and digital video disc (DVD), magnetic storage devices such as magnetic tape and disks, and other non-transferable media used to store information accessible to computing devices. It is important to note that the computer-readable media described herein do not include temporary storage media, such as modulated data signals and carrier waves.

[0066] Those skilled in the art will further recognize that the operation of the module can be achieved using existing technical protocols or programs, without relying on new computer programs themselves. The units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0067] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented in hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0068] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for estimating direction-of-arrival (DOA) of a quantum dream optimization mechanism under impulsive noise, characterized in that, Includes the following steps: Step 1: Establish an array signal receiving model under impulse noise environment and obtain the array received data matrix; Step 2: Calculate the robust covariance matrix based on the array received data matrix; A hybrid kernel correlation entropy covariance matrix is ​​constructed using at least two different kernel functions; the robust covariance matrix is ​​then fused with the hybrid kernel correlation entropy covariance matrix to obtain an enhanced correlation entropy covariance matrix. Step 3: Perform eigenvalue decomposition on the enhanced robust covariance matrix and establish a small snapshot weighted signal subspace fitting equation based on the enhanced hybrid kernel correlation entropy. The objective of the equation is to solve for the angle parameters that minimize the weighted signal subspace fitting criterion. Step 4: The quantum dream optimization algorithm is used to solve the weighted subspace fitting objective function. The individuals in the quantum dream optimization algorithm population are updated in the search space to obtain the optimal solution. Step 5: Set a phase switching threshold to divide the iteration process into an exploration phase and a development phase: When the current iteration count is less than or equal to the phase switching threshold, only the exploration phase is executed, and the position of individuals in the exploration phase is updated based on the local optimum information of their respective subgroups; when the current iteration count is greater than the phase switching threshold but less than the maximum iteration count, only the development phase is executed, and the position of individuals in the development phase is updated based on the global optimum information. Step 6: Determine if the current iteration count has reached the maximum iteration count. If it has, output the angle parameter corresponding to the current global optimal position as the direction of arrival estimation result; otherwise, update the iteration count and return to step 4.

2. The method for estimating direction of arrival (DOA) of a quantum dream optimization mechanism under impulsive noise as described in claim 1, characterized in that, The array signal receiving model described in step 1 is based on a uniform linear array, which contains M array elements with an element spacing of d. It receives incident signals from N far-field narrowband signal sources, with incident angles of each source being... The array signal receiving model received the first... The narrowband signal of the second snapshot is The received data matrix is : in, For the guiding matrix, For the first Sub-sampling signal vector, For the first Sub-sampling impact noise vector This indicates that the signal received by the signal receiving model is the first... The next snapshot is a narrowband signal, where K represents the total number of snapshots. This represents a snapshot narrowband signal vector received by the signal receiving model. This represents the vector of a single snapshot signal. This represents the impulse noise vector of a single snapshot.

3. The method for estimating direction of arrival (DOA) of a quantum dream optimization mechanism under impulsive noise as described in claim 2, characterized in that, The enhanced correlation entropy covariance matrix mentioned in step 2 for: in, These are the weighted combination coefficients. ; Let the robust covariance matrix be denoted as , and its . Line number Column elements are ,in When it satisfies Hermitian symmetry, that is , represent The Middle All elements in the row, represent The Middle All elements in the row, This is an element-wise multiplication operation. To perform conjugate operations, For the median operation, , ; Let the relevant entropy covariance matrix be represented, and its i-th Line number The elements of the column are ,in , , , These are the weighting coefficients. , For matrix The first in Line number Column elements, For matrix The first in Line number Column elements, and is the kernel function of the correlation entropy.

4. The method for estimating direction of arrival (DOA) of a quantum dream optimization mechanism under impulsive noise as described in claim 3, characterized in that, In step 3, eigenvalue decomposition is performed on the enhanced robust covariance matrix to establish an angle estimation equation based on the small snapshot weighted signal subspace fitting equation of the enhanced hybrid kernel correlation entropy. in, The estimated vector of the incoming wave angle that maximizes the projection trace of the weighted signal subspace; It is a diagonal weighted matrix, whose diagonal elements are determined by eigenvalues; To find the trace function; For guiding matrix The mapping matrix; The signal subspace matrix; is the conjugate transpose of the signal subspace matrix; This is the conjugate transpose operator.

5. The method for estimating direction of arrival (DOA) of a quantum dream optimization mechanism under impulsive noise as described in claim 4, characterized in that, In step 4 of the quantum dream optimization algorithm, the first... The second iteration The quantum position of each individual is ,in, , D represents the dimension of the search space; an integer. Let be the number of iterations; the th The second iteration The mapping position corresponding to the quantum position of each individual is: The mapping rule is ,in, , Mapping position of quantum individual The upper limit of dimensional variables, Mapping position of quantum individual Lower bound of the dimensional variable; calculate the dimensional variable. The fitness value of all positions in the quantum population is , among which, the The generation The expression for the fitness value of the mapped position of a quantum individual is: The population is grouped according to the mean of quantum positions, and each subpopulation evolves independently; the quantum position of the quantum individual with the highest fitness within a group is denoted as . As the collective memory of the group; record the iteration to the . The global optimal fitness value up to date is The quantum position of the corresponding individual is the globally optimal quantum position. .

6. The method for estimating direction of arrival (DOA) of a quantum dream optimization mechanism under impulsive noise as described in claim 5, characterized in that, The stage switching threshold in step 5 is When the current iteration number At that time, the exploration phase is executed; when During the development phase, the initial timing command is executed. , This represents the maximum number of iterations. During the exploration phase: for each quantum individual within the group Random selection Each dimension is a "forgotten dimension," and only the location information of the "forgotten dimension" is updated; with adaptive probability. Implement the forgetting supplementation strategy: in the first In each dimension of forgetting, the quantum rotation angle is calculated based on the collective memory. Applying quantum rotation gates to the first The generation Group 1 The quantum position of an individual To supplement memories and obtain the first The generation A quantum individual in the dimension of forgetting The quantum position on ,in, , The exploration phase involves the magnitude of the rotation angle. To supplement the memory coefficients; with preset probabilities Implementing the Dream Sharing Strategy: In the first In each forgetting dimension, the quantum rotation angle is calculated by randomly selecting the quantum position of another quantum individual in the same subgroup, and updated in the same way as the forgetting replenishment strategy, while the position information in other dimensions remains unchanged.

7. The method for estimating direction of arrival (DOA) of a quantum dream optimization mechanism under impulsive noise as described in claim 6, characterized in that, During the development phase, all quantum individuals were randomly selected. The first dimension is the forgetting dimension; only the position information of the forgetting dimension is updated. Calculate the quantum rotation angle based on the globally optimal quantum position in each forgetting dimension. Applying quantum rotation gates to the first The generation The memory of each individual's quantum position is supplemented to obtain the first... The generation The dimension of forgetting of individual quantum entities quantum position ,in, , To develop the magnitude of the rotation angle.

8. The method for estimating direction of arrival (DOA) of a quantum dream optimization mechanism under impulsive noise as described in claim 7, characterized in that, After the first iteration is completed, the number of iterations is obtained. The first generation Individuals The quantum position in the dimensional search space is Calculate the corresponding mapping position and its fitness value ; Calculate the first The maximum fitness value in a quantum population And compare it with the current global optimal fitness value. Compare; if Then update the global highest fitness value and record it. The corresponding quantum position is the globally optimal quantum position; otherwise, it is recorded. The corresponding quantum position is the globally optimal quantum position.

9. A computer 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 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 8.

Images