Optical phased array sidelobe suppression method based on adaptive differential evolution algorithm

Abstract

The invention discloses an optical phased array sidelobe suppression method based on an adaptive differential evolution algorithm. The method comprises the following steps: S1, initializing a model and setting a termination condition; s2, mutation operation is executed, mutation operation based on a self-adaptive mutation operator is carried out on the current population individuals, and population parameter vectors Yi, g + 1 (i = 1, 2,..., NP) after mutation are obtained; s3, performing crossover operation, and performing crossover operation on the varied parameter vectors Yi, g + 1 (i = 1, 2,..., NP) and parameters Xi, g (i = 1, 2,..., NP) of other target vectors to generate sub-individuals Zi, g + 1 (i = 1, 2,..., NP); s4, boundary condition judgment: judging whether the generated sub-individuals exceed a preset boundary condition or not; s5, executing a selection operation; and S6, outputting the current optimal individual. Through the improved adaptive differential evolution algorithm, the phase modulation parameters of the optical phased array split beams are optimized, and the optical phased array emergent beam side lobe energy is suppressed, so that the effective utilization rate of far-field light lobes is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical phased arrays, and in particular relates to an optical phased array sidelobe suppression method based on an adaptive differential evolution algorithm. Background Art

[0002] Optical phased array technology has the advantages of fast response speed, high pointing accuracy, light weight, and miniaturization. It has great development potential in the fields of free-space laser communication, solid-state laser radar, and long-distance three-dimensional imaging. The optical phased array output beam is formed by coherent synthesis of multi-channel beams after being emitted in space. The phased array far-field spot consists of the main lobe and various levels of side lobes. The side lobes will weaken the main lobe energy and affect the overall pointing accuracy of the far-field spot. The smaller the side lobe energy, the more concentrated the main lobe energy, and the higher the pointing accuracy of the far-field spot.

[0003] In the existing methods, the suppression of optical phased array side lobes is mostly achieved by using genetic algorithms. For example, the invention patent with the publication number CN 112465142 A discloses a method for suppressing the side lobes of optical phased array antennas under the framework of genetic algorithms. The method establishes a data set for coordinate optimization by designing a set of antenna arrays arranged in concentric rings, randomly extracts the initial coordinate position from the set according to the required number of antennas, optimizes the suppression of the side lobes of the coordinates of the optical phased array antennas by using the genetic algorithm framework, realizes efficient calculation of the fitness function by fast Fourier transform, designs a chromosome gene point-by-point crossover method to improve the optimization efficiency, and realizes the global optimal light field arrangement scheme by using mutation, etc., which is used to solve the problem of maximizing the energy of the main beam of the optical phased array antenna. However, this method is mainly used to optimize the optical phased array antenna arranged in concentric rings, and cannot be applied to the optical phased array antenna arranged in one-dimensional linear arrangement and other optical phased array antennas arranged in other ways; and the algorithm running time of the genetic algorithm is long when processing large-scale data, and it is difficult to converge to the global optimal point quickly.

[0004] Therefore, it is an urgent technical problem to be solved by technicians in this field to find an algorithm that can suppress the sidelobe energy of optical phased arrays with one-dimensional linear arrangement and other arrangements, reduce the impact of sidelobe on the pointing accuracy of its far-field spot, and converge quickly. Summary of the invention

[0005] The purpose of the present invention is to provide an optical phased array sidelobe suppression method based on an adaptive differential evolution algorithm, optimize the phase modulation parameters of the optical phased array sub-beams, suppress the sidelobe energy of the optical phased array output beam, and thus improve the effective utilization rate of the far-field light lobe.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] An optical phased array sidelobe suppression method based on an adaptive differential evolution algorithm comprises the following steps:

[0008] S1. Initialize the model and set the termination conditions

[0009] Initialize the population size, adaptive mutation operator, crossover operator, and preset values ​​of boundary conditions, and use the optical phased array parameter vector as the first generation population parameter vector X i,1 (i=1,2,…,N P ), and set the termination condition.

[0010] S2. Perform mutation operation

[0011] Perform mutation operation based on adaptive mutation operator on the current population individuals to obtain the population parameter vector Y after mutation i,g+1 (i=1,2,…,N P ).

[0012] S3. Perform crossover operation

[0013] The mutated parameter vector Y i,g+1 (i=1,2,…,N P ) and other target vector parameters X i,g (i=1,2,…,N P ) performs a crossover operation to generate sub-individual Z i,g+1 (i=1,2,…,N P ).

[0014] S4. Boundary condition judgment

[0015] Determine whether the generated sub-individual exceeds the preset boundary conditions, if not, jump to step S5;

[0016] If the boundary conditions are exceeded: If it does not exceed k times the boundary constraint difference (i.e. k(X max -X min )), the individual value is set to the adjacent boundary value; if it exceeds k times the boundary constraint difference, the individual value is replaced by a randomly generated parameter vector in the feasible domain.

[0017] S5. Execute selection operation

[0018] The newly generated child individual Z i,g+1 (i=1,2,…,N P ) and the previous generation individual X i,g (i=1,2,…,N P ), substitute the sidelobe suppression ratio function Φ(X) = I (X) / I ′ (X) ,(i=1,2,3,…,N P), select the individual with the largest sidelobe suppression ratio function value as the next generation; where Φ(X) =

[0019] I (X) / I ′ (X) is the functional mapping relationship between the optical phased array phase modulation parameter vector and the sidelobe suppression ratio, X is the optical phased array phase modulation channel parameter vector, and its vector dimension corresponds to the number of optical phased array channels; I (X) I is the main lobe peak light intensity, which is the maximum light intensity near the main lobe position of the light spot; ′ (X) is the side lobe peak light intensity, which is the maximum light intensity near the main lobe position after removing the main lobe peak light intensity. is the test vector Z i,g+1 Will it become a member of the next generation? The differential evolution algorithm compares the test vector with the target vector X in the current population according to the greedy principle. i,g To obtain a better sidelobe suppression ratio, vectors with larger objective function values ​​will appear in the next generation population. All individuals in the next generation are better or at least as good as the corresponding individuals in the current population.

[0020] S6, executing step S2 to step S5 in a loop, stopping after reaching the termination condition, and outputting the current optimal individual, which is the approximate solution of the optimal solution of the model at this time, so as to obtain a parameter vector that makes the optical phased array sidelobe suppression effect better.

[0021] Preferably, in step S1, during the initialization process, it is assumed that all the initialized populations conform to a uniform probability distribution; let the limit of the parameter variable be X jmin <X j <X jmax , then X ji,0 =rand[0,1]

[0022] (X jmax -X jmin )+X jmin (i=1,2,…,N p ; j = 1, 2, …, d), where rand[0, 1] represents a uniform random number generated between [0, 1].

[0023] Preferably, in step S2, for each target vector X ig (i=1,2,…,N P ), the mutation vector based on the adaptive differential evolution algorithm is as follows: In the formula, the randomly selected numbers r1, r2, and r3 are different from each other, and r1, r2, and r3 are also different from the target vector number i, and must satisfy N P ≥4; Adaptive mutation operator F∈[0,2]; Adaptive mutation operator Where F0 represents the mutation operator, G m represents the maximum evolutionary generation, G represents the current evolutionary generation; at the beginning, the adaptive mutation operator is 2F0, which has a large value, which helps to maintain the diversity of individuals in the population and avoid premature maturity in the early stage; as the algorithm progresses, the mutation operator gradually decreases, and the mutation rate is close to F0 in the later stage, which retains good information, avoids the destruction of the optimal solution, and increases the probability of the global optimal solution.

[0024] Preferably, in step S3,

[0025]

[0026] where randb(j) represents the jth estimated value of a random number generator between [0,1]; rnbr(i)∈(1,2,…,D) represents a randomly selected sequence that ensures that Z i,g+1 At least from Y i,g+1 Get a parameter; CR represents the crossover operator, and its value range is [0,1].

[0027] The concept of the present invention is:

[0028] The optical phased array controls the composite antenna composed of multiple optical channels arranged in a regular pattern to make the light beams of each channel coherently synthesized in space, thereby achieving far-field light beam pointing control. Taking the optical waveguide optical phased array as an example, adjusting the phase modulation parameters of each channel of the optical phased array can independently control the output light phase of each channel and change the sidelobe suppression ratio of the far-field light spot of the optical phased array. A real-time monitoring platform for the far-field light spot intensity of the optical phased array is built, and the main lobe peak light intensity and sidelobe peak light intensity are monitored in real time by an infrared camera to obtain the corresponding relationship between the phase parameter vector of the optical phased array and the main lobe peak light intensity and sidelobe peak light intensity. The phase parameter vector is continuously iterated using an adaptive differential evolution algorithm to solve the parameter vector that makes the sidelobe suppression ratio close to the maximum value, thereby converting the sidelobe suppression problem into a problem of finding a global optimal solution through an intelligent optimization method.

[0029] The real-time monitoring process of the far-field spot light intensity of the optical phased array is as follows:

[0030] The adaptive differential evolution algorithm model outputs the current phase modulation parameter vector to the parameter vector sending module, which synchronously sends the phase modulation parameter vector of each channel to each channel of the optical phased array antenna. The optical phased array outgoing light beam is projected onto a fixed screen, and the infrared camera captures the far-field spot image on the fixed screen and transmits it to the computer to identify the main lobe peak light intensity and side lobe peak light intensity of the optical phased array, calculate the sidelobe suppression ratio corresponding to the parameter vector and input it into the differential evolution algorithm model.

[0031] Furthermore, the sidelobe suppression ratio of the optical phased array is calculated as follows: the infrared camera captures the far-field spot image corresponding to the phase modulation parameter vector X, and the point with the largest light intensity in the spot image is recorded as the main lobe position of the optical phased array. The maximum light intensity near this point is recorded as the main lobe peak light intensity, that is, I (X) , after removing the main lobe peak light intensity, the maximum light intensity near this point is recorded as the side lobe peak light intensity, that is, I ′ (X) , then the functional mapping relationship between the optical phased array phase modulation parameter vector and the optical phased array sidelobe suppression ratio is Φ(X)=I (X) / I ′ (X) , where X is the optical phased array phase modulation channel parameter vector, and its vector dimension corresponds to the number of optical phased array channels.

[0032] Compared with the prior art, the present invention has the following advantages:

[0033] (1) The model parameters can be converged quickly, the phase modulation parameters of the optical phased array beams can be optimized, and the sidelobe energy of the optical phased array output beams can be suppressed, thereby improving the effective utilization of the far-field light lobe and the overall pointing accuracy of the far-field light spot.

[0034] (2) An improved adaptive mutation operator is used to maintain the individual diversity of the population in the early stage and avoid premature maturity; in the later stage, the excellent information of the population is retained to avoid the destruction of the optimal solution and increase the probability of searching for the global optimal solution.

[0035] (3) An improved boundary condition processing strategy is adopted, which can more accurately judge the individual values ​​beyond the boundary, avoid introducing irrelevant individual values, and enable the algorithm to converge faster.

[0036] (4) The present invention adopts the differential evolution algorithm, which is an efficient parallel optimization algorithm. It generates new individuals by random deviation perturbation, guides the search results to approach the global optimal solution, and can quickly iterate to solve the phase modulation parameters of the optical phased array beams. It has strong robustness and global optimization capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is the algorithm flow chart of the present invention.

[0038] Figure 2 It is a schematic diagram of real-time monitoring of the far-field spot light intensity of the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0040] This embodiment adopts a mutation operation based on adaptive differential evolution and a "one-to-one" competitive survival strategy, uses floating-point encoding, and performs optimization calculations in a continuous space. The algorithm flow is as follows: Figure 1 As shown:

[0041] (1) Input the control parameters of the differential evolution algorithm, including the population size N P , mutation operator F, crossover operator CR, maximum evolutionary generation G max , Termination condition G≥G max etc.; set the termination condition to reach the maximum number of evolution generations.

[0042] (2) Randomly generate an initial population, let the individual be X i,g (i=1,2,…,N P ), evolutionary algebra k=1.

[0043] (3) Evaluate the initial population, that is, calculate the sidelobe suppression ratio corresponding to each individual in the initial population

[0044] (4) Based on the adaptive mutation operator, perform mutation operation to obtain Y i,g+1 (i=1,2,…,N P ).

[0045] (5) Select the interference parameter vector and perform crossover operation to obtain Z i,g+1 (i=1,2,…,N P ).

[0046] (6) Determine the boundary conditions. If Z i,g+1 The element value in the vector is greater than or equal to k(X max -X min ), go to step (7), if it is less than k(X max -X min ), go to step (8). max is the upper limit of the boundary value, X min The lower boundary value.

[0047] (7) For values ​​that exceed the boundary, new individuals that meet the boundary constraints are randomly generated to obtain a temporary population.

[0048] (8) For values ​​beyond the boundary, boundary absorption processing is performed, and Z i,g+1 The values ​​of the elements in the vector that exceed the boundary value are equal to the boundary value, and a temporary population is obtained.

[0049] (9) Evaluate the temporary population and calculate the sidelobe suppression ratio of each individual in the temporary population

[0050]

[0051] (10) Perform a one-to-one selection operation on the individuals in the temporary population and the corresponding individuals in the original population, and compare X i,g The corresponding sidelobe suppression ratio Φ(X i,g ) and Z i,g+1 The corresponding sidelobe suppression ratio Φ(Z i,g+1 ), select individuals with larger sidelobe suppression ratio to appear in the next generation population, thus obtaining a new generation population.

[0052] (11) Determine whether the termination condition is met or the maximum evolutionary generation is reached: If so, the evolution terminates and goes to step (12), and the best individual at this time is output as the solution; otherwise, the evolutionary generation k = k + 1, and goes to step (4).

[0053] (12) Output the optimal individual and end.

[0054] In step (4), It is that for each target vector X ig (i=1,2,…,N P ) is obtained based on the mutation vector of the differential evolution algorithm. In the above formula, the randomly selected numbers r1, r2, and r3 are different from each other, and r1, r2, and r3 are also different from the target vector number i, and must satisfy N P ≥4; adaptive mutation operator F∈[0,2]; adaptive mutation operator F=F0×2 τ , Where F0 represents the mutation operator, G m represents the maximum evolutionary generation, G represents the current evolutionary generation; at the beginning, the adaptive mutation operator is 2F0, which has a large value, which helps to maintain the diversity of individuals in the population and avoid premature maturity in the early stage; as the algorithm progresses, the mutation operator gradually decreases, and the mutation rate is close to F0 in the later stage, which retains good information, avoids the destruction of the optimal solution, and increases the probability of the global optimal solution.

[0055] In step (5),

[0056]

[0057] where randb(j) represents the jth estimated value of a random number generator between [0, 1]; rnbr(i)∈(1, 2, …, D) represents a randomly selected sequence that ensures that Z i,g+1 At least from Y i,g+1 Get a parameter; CR represents the crossover operator, and its value range is [0, 1].

[0058] The method for real-time monitoring of the far-field spot light intensity of an optical phased array according to the present invention is as follows:

[0059] The adaptive differential evolution algorithm model outputs the current phase modulation parameter vector to the parameter vector sending module, which synchronously sends the phase modulation parameter vector of each channel to each channel of the optical phased array antenna. The optical phased array outgoing beam is projected onto a fixed screen, and the infrared camera captures the far-field spot image on the fixed screen and transmits it to the computer to identify the main lobe peak light intensity and side lobe peak light intensity of the optical phased array, calculate the side lobe suppression ratio corresponding to the parameter vector and input it into the differential evolution algorithm model, such as Figure 2 shown.

[0060] The above embodiments are only specific examples for further describing the purpose, technical solutions and beneficial effects of the present invention, and the present invention is not limited thereto. Any modification, equivalent replacement, improvement, etc. made within the scope of the disclosure of the present invention are included in the protection scope of the present invention.

Claims

1. An optical phased array sidelobe suppression method based on an adaptive differential evolution algorithm, characterized in that: The following steps are involved: S1. Initialize the model and set the termination conditions Initialize the population size, adaptive mutation operator, crossover operator, and preset values ​​of boundary conditions, and use the optical phased array parameter vector as the first generation population parameter vector X i,1 (i=1,2,…,N P ), and set the termination condition; S2. Perform mutation operation Perform mutation operation based on adaptive mutation operator on the current population individuals to obtain the population parameter vector Y after mutation i,g+1 (i=1,2,…,N P ); S3. Perform crossover operation The mutated parameter vector Y i,g+1 (i=1,2,…,N P ) and other target vector parameters X i,g (i=1,2,…,N P ) performs a crossover operation to generate sub-individual Z i,g+1 (i=1,2,…,N P ); S4. Boundary condition judgment Determine whether the generated sub-individual exceeds the preset boundary conditions, if not, jump to step S5; If the boundary conditions are exceeded: if it does not exceed k times the boundary constraint difference, the individual value is set to the adjacent boundary value; if it exceeds k times the boundary constraint difference, the individual value is replaced by a randomly generated parameter vector in the feasible domain; S5. Execute selection operation The newly generated child individual Z i,g+1 (i=1,2,…,N P ) and the previous generation individual X i,g (i=1,2,…,N P ), substitute the sidelobe suppression ratio function Φ(X) = I (X) / I ′ (X) ,(i=1,2,3,…,N P ), select the individual with the largest sidelobe suppression ratio function value as the next generation; where Φ(X) = I (X) / I ′ (X) is the functional mapping relationship between the phase modulation parameter vector of the optical phased array and the sidelobe suppression ratio, X is the phase modulation parameter vector of the optical phased array, and its vector dimension corresponds to the number of optical phased array channels; I (X) I is the main lobe peak light intensity, which is the maximum light intensity near the main lobe position of the light spot; ′ (X) is the side lobe peak light intensity, which is the maximum light intensity near the main lobe position after removing the main lobe peak light intensity; S6, executing steps S2 to S5 in a loop, stopping after reaching the termination condition, and outputting the current optimal individual, which is the approximate solution of the optimal solution of the model at this time, so as to obtain a parameter vector that makes the optical phased array sidelobe suppression effect best.

2. The optical phased array sidelobe suppression method based on adaptive differential evolution algorithm according to claim 1, characterized in that: In step S2, for each target vector X ig (i=1,2,…,N P ), the mutation vector based on the differential evolution algorithm is as follows: In the formula, the randomly selected numbers r1, r2, and r3 are different from each other, and r1, r2, and r3 are also different from the target vector number i, and must satisfy N P ≥4; adaptive mutation operator F∈[0,2]; adaptive mutation operator F=F0×2 τ , Where F0 represents the mutation operator, G m represents the maximum evolutionary generation, and G represents the current evolutionary generation; at the beginning, the adaptive mutation operator is 2F0, which has a large value, which helps to maintain the diversity of population individuals and avoid premature maturity in the early stage; As the algorithm progresses, the mutation operator gradually decreases, and in the later stage the mutation rate is close to F0, which preserves good information, avoids the destruction of the optimal solution, and increases the probability of the global optimal solution.

3. The optical phased array sidelobe suppression method based on adaptive differential evolution algorithm according to claim 1, characterized in that: In step S3, Where randb(j) represents the jth estimated value of the random number generator between [0, 1]; rnbr(i)∈(1, 2, ..., D) represents a randomly selected sequence, which is used to ensure that Z i,g+1 At least from Y i,g+1 Get a parameter; CR represents the crossover operator, and its value range is [0, 1].

4. The optical phased array sidelobe suppression method based on adaptive differential evolution algorithm according to claim 1, characterized in that: In step S1, during the initialization process, it is assumed that all the initialized populations conform to the uniform probability distribution; let the limit of the parameter variable be X jmin <X j <X jmax , then X ji,0 =rand[0,1]·(X jmax -X jmin )+X jmin (i=1,2,…,N p ; j = 1, 2, ..., d), where rand[0, 1] represents a uniform random number generated between [0, 1].

5. The optical phased array sidelobe suppression method based on adaptive differential evolution algorithm according to claim 1, characterized in that: The real-time monitoring method of the far-field spot light intensity of the optical phased array is: The adaptive differential evolution algorithm model outputs the current phase modulation parameter vector to the parameter vector sending module, which synchronously sends the phase modulation parameter vector of each channel to each channel of the optical phased array antenna. The optical phased array outgoing light beam is projected onto a fixed screen, and the infrared camera captures the far-field spot image on the fixed screen and transmits it to the computer to identify the main lobe peak light intensity and side lobe peak light intensity of the optical phased array, calculate the sidelobe suppression ratio corresponding to the parameter vector and input it into the differential evolution algorithm model.

6. The optical phased array sidelobe suppression method based on adaptive differential evolution algorithm according to claim 1, characterized in that: The calculation method of the sidelobe suppression ratio of the optical phased array is as follows: the infrared camera captures the far-field spot image corresponding to the phase modulation parameter vector X, and the point with the largest light intensity in the spot image is recorded as the main lobe position of the optical phased array. The maximum light intensity near this point is recorded as the main lobe peak light intensity, that is, I (X) , after removing the main lobe peak light intensity, the maximum light intensity near this point is recorded as the side lobe peak light intensity, that is, I ′ (X) , then the function mapping relationship between the optical phased array phase modulation parameter vector and the optical phased array sidelobe suppression ratio is Φ(X)=I (X) / I ′ (X) , where X is the optical phased array phase modulation channel parameter vector, and its vector dimension corresponds to the number of optical phased array channels.

Images