A Method for Optimizing Sidelobe Level of Transmitting Beam of Subarray-Level Sparse Array

Abstract

The present invention discloses a method for optimizing the sidelobe level of the transmitting beam of a subarray-level sparse array. The method includes: obtaining the relevant parameters of a Ka-band array radar; uniformly dividing the entire array into subarrays to obtain the subarray positions and the number of subarrays; initializing the population parameters and performing encoding operations, where the population parameters include the initial population, the current iteration generation, and the maximum iteration number. The initial population includes NP individuals, and the dimension of each individual is the number of all subarrays in the array; processing the current population according to the genetic algorithm to obtain the first optimal individual; processing the current population according to the improved genetic algorithm to obtain the second optimal individual; comparing the first optimal individual and the second optimal individual to obtain the best optimization effect. On the premise of meeting the sparsity rate, the present invention introduces a subarray division method, uniformly divides the original array into multiple subarrays, and then performs DBF according to the pattern multiplication theorem, improving the operation speed.

Description

Technical Field

[0001] The present invention belongs to the technical field of radar, and particularly relates to a method for optimizing the sidelobe level of a subarray-level sparse array transmitting beam. Background Art

[0002] Currently, the problem of optimizing the sidelobe level of a sparse array transmitting beam is a research difficulty in the field of large-scale array signal processing. This is because most of the traditional sparse array transmitting beamforming technologies are studied based on the element-level radar antenna array. However, with the popularization and use of large radar arrays, the defects of the element-level beamforming technology have become increasingly prominent: huge computational complexity, high software and hardware costs for system implementation, etc. Therefore, first dividing a large-scale radar antenna array into several smaller subarrays and performing digital beamforming by adjusting the positions of the sparse subarrays is a fast and effective solution. In this way, by using the output information of a single small subarray, the overall computational complexity can be effectively reduced, the convergence speed can be accelerated, and the software and hardware costs can be reduced to a great extent.

[0003] The genetic algorithm is to minimize the maximum relative sidelobe level of the array by appropriately selecting some subarrays on the premise of a given array geometry, array sparsity rate, and subarray shape.

[0004] However, the standard genetic algorithm has a slow iteration speed and only a slight improvement in the iteration effect, so the optimization result of the sidelobe level of the subarray-level array beam is poor. Summary of the Invention

[0005] In order to solve the above problems existing in the prior art, the present invention provides a method for optimizing the sidelobe level of a subarray-level sparse array transmitting beam. The technical problems to be solved by the present invention are realized through the following technical solutions:

[0006] A method for optimizing the sidelobe level of a subarray-level sparse array transmitting beam, the optimization method comprising:

[0007] Step 1, obtaining relevant parameters of the Ka-band array radar, where the relevant parameters include aperture, operating frequency, the scale of each subarray, and the sparsity rate of the array;

[0008] Step 2, uniformly dividing the entire array into subarrays to obtain the subarray positions and the number of subarrays, and determining the beam directions within the subarrays and between the subarrays;

[0009] Step 3, initializing the population parameters and performing encoding operations, where the population parameters include the initial population, the current iteration generation, and the maximum iteration number, the initial population includes NP individuals, and the dimension of each individual is the number of all subarrays in the array;

[0010] Step 4: Process the current population according to the genetic algorithm to obtain the first optimal individual;

[0011] Step 5: Process the current population according to the improved genetic algorithm to obtain the second optimal individual;

[0012] Step 6: Compare the first optimal individual and the second optimal individual to obtain the best optimization effect.

[0013] In an embodiment of the present invention, uniform subarray partitioning is performed on the entire array to obtain the subarray positions and the number of subarrays, including:

[0014] When the element spacing is half a wavelength, the entire array is uniformly partitioned into subarrays using the aperture and the scale of each subarray to obtain the subarray positions and the number of subarrays.

[0015] In an embodiment of the present invention, the initial parameter values of the individuals are:

[0016] f ji,0 = randn[0,1], i = 1, 2, …, NP; j = 1, 2, …, L

[0017] In the formula, randn[0,1] represents a random number that conforms to the Gaussian distribution generated between [0,1]. Let the values of the largest NL genes in each individual be 1, and the values of the remaining genes be 0. L is the number of genes in each individual, and NL is the number of subarrays after sparsification.

[0018] In an embodiment of the present invention, the step 4 includes:

[0019] Step 4.1: Use the roulette wheel selection method for all individuals in the current population to obtain the first population, and the first population includes several individuals for crossover;

[0020] Step 4.2: Perform a crossover operation on the individuals in the odd positions and the individuals in the even positions adjacent to each other in the first population with a first crossover probability Pc to obtain the second population, and the second population includes several individuals for mutation;

[0021] Step 4.3: Based on the first mutation probability Pm, perform a mutation operation on the individuals in the second population to obtain the third population, and the third population includes several mutated individuals;

[0022] Step 4.4: Determine the relationship between the number of genes of the individuals in the third population and NL. If the number of genes of the individuals in the third population is greater than NL, randomly select a preset number of genes from the genes with a status of 1, and set the working status of these genes to 0 to obtain a fourth population. If the number of genes of the individuals in the third population is less than NL, randomly select a preset number of genes from the genes with a status of 0, and set the working status of these genes to 1 to obtain a fourth population. Here, NL is the number of sub-arrays after sparsification, and the preset number is the difference between the number of genes of the individuals in the third population and NL.

[0023] Step 4.5: Determine whether the current iteration number has reached the maximum iteration number. If not, loop and execute Steps 4.1 to 4.5. If so, take the individual with the maximum fitness value in the fourth population as the first optimal individual.

[0024] In an embodiment of the present invention, Step 4.2 includes:

[0025] Take the individuals at odd positions and even positions adjacent to each other in the first population as a pair of paired individuals to be mated. Randomly select K1 positions with a value of 1 from the first sequence [1, L - 1] as the crossover positions. Generate a random number r1 in the interval [0, 1]. If the random number r1 is less than the first crossover probability Pc, the paired individuals exchange their respective partial genes at the crossover positions to obtain new individuals. All the new individuals form the second population, where L is the bit string length.

[0026] In an embodiment of the present invention, Step 4.3 includes:

[0027] In the second population, generate a random number r2 in the interval [0, 1]. If the random number r2 is less than the first mutation probability Pm, take the (j, i)-th gene as the mutant gene. If the gene value of the mutant gene is 1, its gene value becomes 0. If the gene value of the mutant gene is 0, its gene value becomes 1 to obtain new individuals. All the new individuals form the third population, where j = 1 to N, i = 1 to NP, and N is the number of digits in the individual.

[0028] In an embodiment of the present invention, Step 5 includes:

[0029] Step 5.1: Select the individual with the maximum fitness value from all the individuals in the current population as the individual for crossover.

[0030] Step 5.2: Use the individuals selected for crossover in Step 5.1 to perform a crossover operation with all the individuals at even positions in the current population at the second crossover probability Pc to obtain a fifth population, where the fifth population includes several individuals for mutation;

[0031] Step 5.3: Based on the second mutation probability Pm, perform a multi-point mutation operation on several genes of all individuals in the fifth population to obtain a sixth population;

[0032] Step 5.4: Sort the fitness values of the individuals in the sixth population and the individuals in the parental population in Step 5.1, and select the top NP individuals with larger fitness values to form a seventh population;

[0033] Step 5.5: Judge the relationship between the number of array elements of the individuals in the seventh population and NL. If the number of array elements of the individuals in the seventh population is greater than NL, randomly select a preset number of array elements from the array elements with a state of 1 and set the working state of this array element to 0 to obtain an eighth population. If the number of array elements of the individuals in the seventh population is less than NL, randomly select a preset number of array elements from the array elements with a state of 0 and set the working state of this array element to 1 to obtain an eighth population, where NL is the number of sub-arrays after sparsification, and the preset number is the difference between the number of array elements of the individuals in the seventh population and NL;

[0034] Step 5.6: Judge whether the current iteration number reaches the maximum iteration number. If not, loop and execute Steps 5.1 to 5.6. If so, use the individual with the largest fitness value in the eighth population as the second optimal individual.

[0035] In an embodiment of the present invention, Step 5.2 includes:

[0036] Use the individuals selected for crossover in Step 5.1 and the individuals at even positions in the current population as a pair of mating individuals. Randomly select K2 positions with a value of 1 from the second sequence [1, L - 1] as the crossover point positions. Generate a random number r3 in the interval [0, 1]. If the random number r3 is less than the second crossover probability Pc, the mating individuals exchange their respective partial genes at the crossover positions to obtain new individuals, and all the new individuals form the fifth population, where L is the bit string length.

[0037] In an embodiment of the present invention, Step 5.4 includes:

[0038] Generate a random number r2 in the interval [0, 1]. If the random number r2 is less than the second mutation probability Pm, perform mutation operations on some genes randomly selected from the individuals in the fifth population. If the gene value of a gene is 1, its gene value is changed to 0; if the gene value of a gene is 0, its gene value is changed to 1, so as to obtain new individuals, and all the new individuals form the sixth population.

[0039] In an embodiment of the present invention, the fitness value is calculated through a fitness function, and the fitness function is:

[0040]

[0041] where max represents the maximum value function, S1 represents the sidelobe interval of the azimuth pattern when θ = θ0, S2 represents the sidelobe interval of the azimuth pattern when θ represents the elevation angle, represents the azimuth angle, θ0 represents the expected direction in the elevation dimension, represents the pattern function within the subarray, represents the pattern function between the subarrays.

[0042] Advantages of the present invention:

[0043] On the premise of meeting the sparsity ratio, the present invention introduces a subarray division method, evenly divides the original array into multiple subarrays, and then performs DBF according to the pattern multiplication theorem, improving the operation speed.

[0044] The method for optimizing the sidelobe level of the emission beam of the subarray-level sparse array provided by the present invention uses the "monarch scheme", that is, it is improved on the selection and crossover operations of the standard genetic algorithm. Therefore, while ensuring that the array aperture remains unchanged, the optimization result is greatly improved.

[0045] The following will further elaborate on the present invention in conjunction with the drawings and embodiments. Description of the Drawings

[0046] Figure 1 is a schematic flowchart of a method for optimizing the sidelobe level of the emission beam of a subarray-level sparse array provided by an embodiment of the present invention;

[0047] Figure 2 is a flowchart of a method for optimizing the sidelobe level of the emission beam of a subarray-level sparse array provided by an embodiment of the present invention;

[0048] Figure 3 is the result of subarray division under a given array aperture provided by an embodiment of the present invention;

[0049] Figure 4 is the transmit beam pattern after two-dimensional DBF in the full array case provided by the embodiment of the present invention;

[0050] Figure 5 is the sub-array distribution diagram after randomly selecting sub-arrays under a given sparsity rate provided by the embodiment of the present invention;

[0051] Figure 6 is the transmit beam pattern after two-dimensional DBF after randomly sparsifying the array provided by the embodiment of the present invention;

[0052] Figure 7 is the sub-array distribution diagram after optimizing the sub-array positions using the genetic (GA) algorithm under a given sparsity rate provided by the embodiment of the present invention;

[0053] Figure 8 is the transmit beam pattern after two-dimensional DBF after sparsifying the array by GA provided by the embodiment of the present invention;

[0054] Figure 9 is the sub-array distribution diagram after optimizing the sub-array positions using an improved genetic algorithm under a given sparsity rate provided by the embodiment of the present invention;

[0055] Figure 10 is the transmit beam pattern after two-dimensional DBF after sparsifying the array using the improved GA algorithm provided by the embodiment of the present invention;

[0056] Figure 11 is the elevation dimension beam comparison diagram obtained by slicing the array transmit beam patterns obtained under different conditions in the desired direction provided by the embodiment of the present invention;

[0057] Figure 12 is the azimuth dimension beam comparison diagram obtained by slicing the array transmit beam patterns obtained under different conditions in the desired direction provided by the embodiment of the present invention;

[0058] Figure 13 is the comparison diagram of the fitness evolution curves of the standard genetic algorithm and the improved genetic algorithm provided by the embodiment of the present invention. Specific Embodiments

[0059] The following further describes the present invention in detail with specific embodiments, but the embodiments of the present invention are not limited thereto.

[0060] Embodiment 1

[0061] Please refer to Figure 1 and Figure 2 , Figure 1It is a schematic flow chart of a method for optimizing the sidelobe level of a subarray-level sparse array transmit beam provided by an embodiment of the present invention. Figure 2 It is a flow chart of a method for optimizing the sidelobe level of a subarray-level sparse array transmit beam provided by an embodiment of the present invention. An embodiment of the present invention provides a method for optimizing the sidelobe level of a subarray-level sparse array transmit beam. The method for optimizing the sidelobe level of the subarray-level sparse array transmit beam includes steps 1 to 6, where:

[0062] Step 1: Obtain the relevant parameters of the Ka-band array radar. The relevant parameters include the aperture, operating frequency, scale of each subarray, and sparsity rate of the array.

[0063] Step 2: Perform uniform subarray partitioning on the entire array to obtain the subarray positions and the number of subarrays, and determine the beam directions within the subarrays and between the subarrays.

[0064] Specifically, when the element spacing is half a wavelength, use the aperture, the spacing between element positions, and the scale of each subarray to perform uniform subarray partitioning on the entire array to obtain the subarray positions and the number of subarrays.

[0065] For example, please refer to Figure 3 , there are 1764 array elements. Assuming the subarray scale is 3*3, then it can be divided into 196 subarrays, and each subarray has 9 array elements.

[0066] Step 3: Initialize the population parameters and perform the encoding operation. Among them, the population parameters include the initial population, the current iteration generation, and the maximum iteration number. The initial population includes NP individuals, and the dimension of each individual is the number of all subarrays in the array.

[0067] Specifically, initialize the population parameters and perform the encoding operation. Set the evolution generation counter g = 0, set the maximum iteration number to G, generate a binary initial population that meets a certain sparsity rate, randomly generate NP individuals as the initial population, and the dimension of each individual is the number of all subarrays in the array, where each individual is represented as f i,g (i = 1, 2,..., NP), i represents the serial number of the individual in the corresponding population, g represents the genetic generation, and NP represents the number of individuals in a population.

[0068] The individuals of the population need to be initially encoded to establish the initial point of the optimized search. Assume that the number of subarrays after sparsification is NL. Assume that all randomly initialized populations conform to the Gaussian distribution. Then the initial parameter values of the individuals can be obtained by the following formula:

[0069] f ji,0 = randn[0,1], i = 1, 2,..., NP; j = 1, 2,..., L

[0070] In the formula, randn[0,1] represents a random number that follows a Gaussian distribution generated between [0,1]. Let the values of the largest NL genes (i.e., sub-arrays) in each individual be 1, and the values of the remaining genes be 0. L is the number of genes in each individual.

[0071] Calculate the fitness of each individual in this population. First, calculate the two-dimensional emission beam pattern of the sub-array level sparse planar array. The two-dimensional emission beam pattern of the sub-array level sparse planar array can be regarded as the result of the product of the pattern on each individual smaller-scale sub-array divided from the large-scale overall antenna array and the pattern between the sub-arrays. The pattern function within the sub-array is:

[0072]

[0073] The pattern function between the sub-arrays is:

[0074]

[0075] Among them, f mn represents the working state of the corresponding array element, and f mn = 1 indicates that there is a sub-array at the corresponding position; f mn = 0 indicates that there is no sub-array at the corresponding position.

[0076] Therefore, the final pattern function is

[0077] According to the definition of the maximum sidelobe level (MSLL), take the fitness function as the sum of the maximum sidelobe level of the azimuth dimension pattern and the maximum sidelobe level of the elevation dimension pattern, that is, the fitness function is:

[0078]

[0079] In the formula, max represents the maximum value function, S1 represents the sidelobe interval of the azimuth dimension pattern when θ = θ0, S2 represents the sidelobe interval of the azimuth dimension pattern when, θ represents the elevation angle, represents the azimuth angle, θ0 represents the expected direction in the elevation dimension, represents the expected direction in the azimuth angle, represents the pattern function within the said sub-array, represents the pattern function between the said sub-arrays.

[0080] The following optimization model can be defined: (indicating to find the minimum value in MSLL), optimize the sub-array position by optimizing the value of f to minimize the maximum sidelobe level.

[0081] Step 4: Process the current population according to the genetic algorithm to obtain the first optimal individual.

[0082] Step 4.1: Use the roulette wheel selection method for all individuals in the current population to obtain the first population, which includes several individuals for crossover.

[0083] Specifically, the roulette wheel selection method is adopted, and the proportion of the fitness values of each individual is used to determine the possibility of retaining its offspring. If the fitness value of an individual is fit i , and the population size is NP, then the probability of its being selected is expressed as: The larger the individual fitness value, the greater the chance of its being selected; vice versa. To select crossover individuals, multiple rounds of selection are required. For example, select for half the number of rounds of the number of individuals. Generate a uniform random number within [0, 1] in each round, and compare this random number with the probability p i . If the random number is greater than the probability p i , then select this individual; if it is less, then do not select.

[0084] Step 4.2: Perform a crossover operation on the individuals at odd and even positions adjacent to each other in the first population with a first crossover probability Pc to obtain the second population, which includes several individuals for mutation.

[0085] Specifically, first, the individuals at odd and even positions adjacent to each other in the first population are used as a pair of mating individuals. For example, the first individual and the second individual are paired, the third individual and the fourth individual are paired, and so on; then, randomly select K1 positions with a value of 1 from the first sequence [1, L - 1] as the crossover point positions, where L is the bit string length. Since the length of the first sequence is the same as the number of genes in an individual, that is, L is the number of genes in each individual. This first sequence is a sequence composed of 0 and 1, and the value of each position is randomly generated. K1 is the number of positions with a value of 1 in the first sequence; finally, generate a random number r1 in the interval [0, 1]. If the random number r1 is less than the first crossover probability Pc, then the paired individuals exchange their respective partial genes at the crossover positions, that is, swap the genes of the two paired individuals at the crossover positions to obtain new individuals, and all the new individuals form the second population.

[0086] Step 4.3: Based on the first mutation probability Pm, perform a mutation operation on the individuals in the second population to obtain the third population, which includes several mutated individuals.

[0087] Specifically, in the second population, a random number r2 is generated in the interval [0, 1]. If the random number r2 is less than the first mutation probability Pm, the (j, i)-th gene x(j, i) is taken as the mutant gene. If the gene value of the mutant gene is 1, its gene value is changed to 0; if the gene value of the mutant gene is 0, its gene value is changed to 1 to obtain new individuals. All the new individuals form the third population, where j = 1 to N, i = 1 to NP, and N is the number of digits in an individual.

[0088] Step 4.4: To ensure that the sparsity rate of each individual in the newly generated population remains unchanged, that is, the number of 0s and 1s remains unchanged, the relationship between the number of genes of the individuals in the third population and NL is judged. If the number of genes of the individuals in the third population is greater than NL, a preset number of genes are randomly selected from the genes with a state of 1, and the working state of this gene is set to 0 to obtain the fourth population. If the number of genes of the individuals in the third population is less than NL, a preset number of genes are randomly selected from the genes with a state of 0, and the working state of this gene is set to 1 to obtain the fourth population, where NL is the number of sub-arrays after sparsification, and the preset number is the difference between the number of genes of the individuals in the third population and NL.

[0089] Step 4.5: Judge whether the current iteration number has reached the maximum iteration number. If not, loop and execute Steps 4.1 to 4.5 until the maximum iteration number is reached. If so, take the individual with the largest fitness value in the fourth population as the first optimal individual.

[0090] Step 5: Process the current population according to the improved genetic algorithm to obtain the second optimal individual.

[0091] Step 5.1: Select the individual with the largest fitness value from all the individuals in the current population as the individual for crossover.

[0092] Step 5.2: Perform a crossover operation on the individual selected in Step 5.1 and all the other individuals at even positions in the current population with the second crossover probability Pc to obtain the fifth population, and the fifth population includes several individuals for mutation.

[0093] Specifically, first, the individuals selected for crossover in step 5.1 and the other individuals with even positions in the current population are used as a pair of mating individuals; then, randomly select K2 positions with a value of 1 from the second sequence [1, L - 1] as the crossover point positions, where the second sequence is a sequence composed of 0s and 1s, and the value of each position is randomly generated, and K2 is the number of positions with a value of 1 in the second sequence; finally, generate a random number r3 in the interval [0, 1]. If the random number r3 is less than the second crossover probability Pc, the paired individuals exchange their respective partial genes at the crossover positions, that is, the genes of the two paired individuals at the crossover positions are swapped to obtain new individuals, and all the new individuals form the fifth population.

[0094] Step 5.3: Based on the second mutation probability Pm, perform a multi-point mutation operation on several genes of all individuals in the fifth population to obtain the sixth population.

[0095] Specifically, generate a random number r2 in the interval [0, 1]. If the random number r2 is less than the second mutation probability Pm, perform a mutation operation on some randomly selected genes in the individuals in the fifth population. If the gene value of a gene is 1, its gene value becomes 0, and if the gene value of a gene is 0, its gene value becomes 1 to obtain new individuals, and all the new individuals form the sixth population. For example, each individual has 50 genes in total, and 10 genes are randomly selected for mutation operation.

[0096] Step 5.4: Combine the individuals in the sixth population with the parental population in step 5.1, and sort the fitness values of the combined individuals, and select the top NP individuals with larger fitness values to form the seventh population.

[0097] Step 5.5: To ensure that the sparsity rate of each individual in the newly generated population remains unchanged, that is, the number of 0s and 1s remains unchanged, then judge the relationship between the number of genes of the individuals in the seventh population and NL. If the number of genes of the individuals in the seventh population is greater than NL, randomly select a preset number of genes from the genes with a state of 1 and set the working state of the gene to 0 to obtain the eighth population. If the number of genes of the individuals in the seventh population is less than NL, randomly select a preset number of genes from the genes with a state of 0 and set the working state of the gene to 1 to obtain the eighth population, where NL is the number of sub-arrays after sparsification, and the preset number is the difference between the number of genes of the individuals in the seventh population and NL.

[0098] Step 5.6: Judge whether the current iteration number has reached the maximum iteration number. If not, loop and execute steps 5.1 to 5.6 until the maximum iteration number is reached. If so, take the individual with the largest fitness value in the eighth population as the second optimal individual.

[0099] Step 6: Compare the first optimal individual with the second optimal individual to obtain the best optimization effect.

[0100] Specifically, according to the two results obtained in Step 4 and Step 5, considering the non-sparse (i.e., full matrix) and randomly sparse cases, the comparison results of the transmitting beam direction diagrams in the azimuth dimension and elevation dimension are respectively drawn for selection. The results show that the improved genetic algorithm has better iteration effects.

[0101] In summary, the embodiment of the present invention improves the optimization process of the sidelobe level of the Ka-band subarray-level sparse planar array, and can be applied to general medium and small-scale sparse array radar systems.

[0102] To verify the effectiveness of the subarray-level sparse array transmitting beam sidelobe level optimization method based on the improved genetic algorithm provided by the embodiment of the present invention, the inventor conducted a simulation experiment, and the details of the simulation experiment are further described below:

[0103] (1) Simulation conditions: First, describe the simulation conditions of the improved genetic algorithm, including the number of individuals being 40, the crossover probability being 0.8, the mutation probability being 0.1, and the number of iterations being 500 times. Second, describe the simulation conditions of the planar array, including the operating frequency being 33 GHz, the size of the planar array being 0.2 m * 0.2 m, the array elements being distributed at half-wavelength intervals, each subarray having 3 * 3 array elements, the array sparsity rate being 40%, the two-dimensional beam pointing within and between subarrays both being (0°, 0°), and the scanning ranges of the array in the azimuth dimension and elevation dimension both being -40° to 40°. In addition, the generation and processing of data during the experiment are all completed on the MATLAB software version 2016a. A total of four scenarios are simulated, and the details are as follows:

[0104] Experiment scenario 1: The number of array elements is 1764, the array type is a planar array, the operating wavelength λ is 0.0091 meters, the element spacing d is half-wavelength. According to the set conditions, the planar array can be divided into 196 subarrays, each subarray having 9 array elements, and the scanning ranges of the array in the elevation dimension and azimuth dimension are both -40° to 40°.

[0105] For the simulation results of Experiment scenario 1, see Figure 3 and Figure 4 as shown, Figure 3 which shows the placement position of the array in the embodiment of the present invention, that is, the array is placed in the yoz plane and can be scanned in the elevation dimension and azimuth dimension. Among them, the abscissa represents the y-axis, and the ordinate represents the z-axis. Figure 4 shows the subarray-level two-dimensional DBF of the array in the non-sparse case in the embodiment of the present invention, and this direction diagram can form a peak in the desired direction.

[0106] Experiment scenario 2: On the basis of Experiment scenario 1, the planar array is sparsified with a sparsity rate of 40%, and at this time, the selection of subarrays is randomly selected.

[0107] For the simulation results of Experimental Scenario 2, see Figure 5 and Figure 6 as shown in Figure 5 which shows the sub - array distribution diagram after random sparsity in the embodiments of the present invention, Figure 6 and which shows the two - dimensional sub - array - level DBF direction diagram after random sparsity in the embodiments of the present invention. By comparing with Figure 4 it can be seen that the overall sidelobe level of the direction diagram has increased, and the sidelobe level at (0°, 0°) has increased a lot.

[0108] Experimental Scenario 3: On the basis of Experimental Scenario 1, the planar array is sparsified and optimized using the standard genetic algorithm, with 500 iterations and a sparsity rate of 40%.

[0109] For the simulation results of Experimental Scenario 3, see Figure 7 and Figure 8 as shown in Figure 7 which shows the sub - array position distribution diagram after optimization by the standard genetic algorithm in the embodiments of the present invention, Figure 8 and which shows the two - dimensional sub - array - level DBF direction diagram after optimization by the standard genetic algorithm. By comparing with Figure 6 it can be seen that the sidelobe level of the direction diagram has decreased a lot, but there is still a large room for optimization.

[0110] Experimental Scenario 4: On the basis of Experimental Scenario 1, the planar array is sparsified and optimized using the improved genetic algorithm, with 500 iterations and a sparsity rate of 40%.

[0111] For the simulation results of Experimental Scenario 4, see Figure 9 and Figure 10 as shown in Figure 9 which shows the sub - array position distribution diagram after optimization by the standard genetic algorithm in the embodiments of the present invention, Figure 10 and which shows the two - dimensional sub - array - level DBF direction diagram after optimization by the standard genetic algorithm. By comparing with Figure 8 it can be seen that the sidelobe level of the direction diagram has been further reduced.

[0112] Figure 11 and Figure 12 give the comparison results of the direction diagrams of the four experimental scenarios in the azimuth dimension and the elevation dimension. It can be seen that, whether in the azimuth dimension or the elevation dimension, the method provided by the embodiments of the present invention can achieve very good results in optimizing the sidelobe level. And through the comparison of the fitness evolution curves in Figure 13 it can be seen that the embodiments of the present invention are greatly optimized in the calculation process, improving the optimization process of the sidelobe level of the transmitting beam direction diagram of the Ka - band sub - array - level planar array radar.

[0113] In traditional sparse array transmission DBF (Digital Beam Forming) technology, the vast majority of research is based on the array element-level radar antenna array. With the popularization and use of large radar arrays, the defects of the array element-level DBF technology have become prominent, such as its huge amount of computation and high system implementation software and hardware costs, which seriously affect the sidelobe level optimization process. Under the premise of meeting the sparsity rate, the present invention introduces a sub-array division method, evenly divides the original array into multiple sub-arrays, and then performs DBF according to the pattern multiplication theorem, improving the operation speed.

[0114] When a certain sparsity rate is satisfied, if the sub-array positions are randomly selected, it will inevitably result in a relatively high sidelobe level. However, if the standard genetic algorithm is used, its iterative effect is poor. The method for optimizing the sidelobe level of the sub-array-level sparse array transmission beam provided by the present invention uses the "monarch scheme", that is, it is improved in the selection and crossover operations of the standard genetic algorithm. Therefore, while ensuring that the array aperture remains unchanged, the optimization result has been greatly improved.

[0115] Those of ordinary skill in the art can understand that all or part of the steps for implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above method embodiments; and the foregoing storage medium includes: various media such as ROM, RAM, magnetic disk, or optical disc that can store program codes.

[0116] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine the different embodiments or examples described in this specification.

[0117] Although the present application has been described in connection with various embodiments, however, in the process of implementing the claimed present application, those skilled in the art can understand and achieve other variations of the disclosed embodiments by viewing the accompanying drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "one" does not exclude a plurality. A single processor or other unit may implement several functions recited in the claims. Certain measures are recited in mutually different dependent claims, but this does not mean that these measures cannot be combined to produce good results.

[0118] The above content is a further detailed description of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention pertains, without departing from the concept of the present invention, several simple deductions or substitutions can still be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. A method for optimizing the sidelobe level of the transmitting beam of a subarray-level sparse array, characterized in that The optimization method includes: Step 1: Obtain the relevant parameters of the Ka-band array radar, where the relevant parameters include aperture, operating frequency, the scale of each sub-array, and the sparsity rate of the array; Step 2: Divide the entire array into uniform sub-arrays to obtain the sub-array positions and the number of sub-arrays, and determine the beam directions within the sub-arrays and between the sub-arrays; Step 3: Initialize the population parameters and perform encoding operations. Among them, the population parameters include the initial population, the current iteration generation, and the maximum iteration number. The initial population includes NP individuals, and the dimension of each individual is the number of all sub-arrays in the array; Step 4: Process the current population according to the genetic algorithm to obtain the first optimal individual; Step 5: Process the current population according to the improved genetic algorithm to obtain the second optimal individual; Step 6: Compare the first optimal individual and the second optimal individual to obtain the best optimization effect; The step 4 includes: Step 4.1: Use the roulette wheel selection method for all individuals in the current population to obtain the first population, and the first population includes several individuals for crossover; Step 4.2: Perform crossover operations on the individuals in the odd positions and the even positions adjacent to each other in the first population with a first crossover probability Pc to obtain the second population, and the second population includes several individuals for mutation; Step 4.3: Based on the first mutation probability Pm, perform mutation operations on the individuals in the second population to obtain the third population, and the third population includes several mutated individuals; Step 4.4: Judge the relationship between the number of genes of the individuals in the third population and NL. If the number of genes of the individuals in the third population is greater than NL, randomly select a preset number of genes from the genes with a state of 1 and set the working state of this gene to 0 to obtain the fourth population. If the number of genes of the individuals in the third population is less than NL, randomly select a preset number of genes from the genes with a state of 0 and set the working state of this gene to 1 to obtain the fourth population, where NL is the number of sub-arrays after sparsification, and the preset number is the difference between the number of genes of the individuals in the third population and NL; Step 4.5: Judge whether the current iteration number reaches the maximum iteration number. If not, loop and execute steps 4.1 to 4.

5. If so, use the individual with the largest fitness value in the fourth population as the first optimal individual; The step 5 includes: Step 5.1: Select the individual with the largest fitness value from all individuals in the current population as the individual for crossover; Step 5.2: Use the individual for crossover selected in step 5.1 to perform crossover operations with all other even-position individuals in the current population with a second crossover probability Pc to obtain the fifth population, and the fifth population includes several individuals for mutation; Step 5.3: Based on the second mutation probability Pm, perform multi-point mutation operations on several genes of all individuals in the fifth population to obtain the sixth population; Step 5.4: Sort the individuals of the sixth population and the individuals of the parental population in step 5.1 according to their fitness values, and select the top NP individuals with larger fitness values to form the seventh population; Step 5.5: Judge the relationship between the number of array elements of the individuals in the seventh population and NL. If the number of array elements of the individuals in the seventh population is greater than NL, randomly select a preset number of array elements from the array elements with state 1, and set the working state of this array element to 0 to obtain the eighth population. If the number of array elements of the individuals in the seventh population is less than NL, randomly select a preset number of array elements from the array elements with state 0, and set the working state of this array element to 1 to obtain the eighth population, where NL is the number of sub-arrays after sparsification, and the preset number is the difference between the number of array elements of the individuals in the seventh population and NL; Step 5.6: Judge whether the current iteration number reaches the maximum iteration number. If not, loop and execute steps 5.1 to 5.

6. If so, take the individual with the largest fitness value in the eighth population as the second optimal individual; The fitness value is calculated by a fitness function, and the fitness function is: Among them, max represents the maximum value function, S1 represents the sidelobe interval of the azimuth pattern when θ = θ0, and S2 represents the sidelobe interval of the azimuth pattern when, θ represents the elevation angle, represents the azimuth angle, θ0 represents the desired direction in the elevation dimension, represents the desired direction in the azimuth angle, represents the pattern function within the subarray, represents the pattern function between the subarrays.

2. The method for optimizing the sidelobe level of the transmitting beam of the subarray-level sparse array according to claim 1, wherein Perform uniform sub-array partitioning on the entire array to obtain the sub-array positions and the number of sub-arrays, including: When the element spacing is half a wavelength, use the aperture and the scale of each sub-array to perform uniform sub-array partitioning on the entire array to obtain the sub-array positions and the number of sub-arrays.

3. The method for optimizing the sidelobe level of the transmitting beam of the subarray-level sparse array according to claim 1, characterized in that The initial parameter values of the individual are: f ji,0 = randn[0,1], i = 1, 2, …, NP; j = 1, 2, …, L In the formula, randn[0,1] represents a random number that conforms to the Gaussian distribution generated between [0,1]. Let the values of the largest NL genes in each individual be 1, and the values of the remaining genes be 0. L is the number of genes of each individual, and NL is the number of sub-arrays after sparsification.

4. The method for optimizing the sidelobe level of the sub-array level sparse array transmission beam according to claim 1, wherein The step 4.2 includes: Take the individuals at adjacent odd positions and even positions in the first population as a pair of paired individuals to be mated. Randomly select K1 positions where the integer is 1 from the first sequence [1, L-1] as the crossover point positions. Generate a random number r1 in the interval [0,1]. If the random number r1 is less than the first crossover probability Pc, the paired individuals exchange their respective partial genes at the crossover positions to obtain new individuals, and all the new individuals form the second population. L is the bit string length.

5. The method for optimizing the sidelobe level of the transmitting beam of the subarray-level sparse array according to claim 1, wherein The step 4.3 includes: In the second population, generate a random number r2 in the interval [0,1]. If the random number r2 is less than the first mutation probability Pm, take the (j,i) gene as the mutated gene. If the gene value of the mutated gene is 1, its gene value becomes 0. If the gene value of the mutated gene is 0, its gene value becomes 1 to obtain new individuals, and all the new individuals form the third population, where j = 1~N, i = 1~NP, and N is the number of digits in the individual.

6. The method for optimizing the sidelobe level of the emission beam of the sub-array level sparse array according to claim 1, wherein The step 5.2 includes: Use the individuals selected for crossover in step 5.1 and the individuals at even positions in the current population as a pair of mating individuals. Randomly select K2 positions with a value of 1 from the second sequence [1, L - 1] as the crossover point positions. Generate a random number r3 in the interval [0, 1]. If the random number r3 is less than the second crossover probability Pc, the mating individuals exchange their respective partial genes at the crossover positions to obtain new individuals. All the new individuals form the fifth population, where L is the bit string length.

7. The method for optimizing the sidelobe level of the transmitting beam of the sub-array level sparse array according to claim 1, wherein Step 5.4 includes: Generate a random number r2 in the interval [0, 1]. If the random number r2 is less than the second mutation probability Pm, perform a mutation operation on some genes randomly selected from the individuals in the fifth population. If the gene value of a gene is 1, its gene value becomes 0; if the gene value of a gene is 0, its gene value becomes 1, to obtain new individuals. All the new individuals form the sixth population.

Images