Filter multi-dimensional coefficient joint optimization method based on improved dung beetle algorithm

Abstract

The invention discloses a filter multi-dimensional coefficient joint optimization method based on an improved dung beetle algorithm, and belongs to the field of filter optimization design, and the method comprises the steps: building an optimization model according to a filter needing to be optimized, and obtaining a to-be-optimized parameter and an optimization objective function; constructing a filter multi-dimensional coefficient joint optimization method based on an improved dung beetle algorithm; and introducing the optimized parameters into a filter, simulating to generate a passive intermodulation component, and performing interference suppression. According to the method, the optimal coefficient of the FIR filter can be efficiently and accurately searched, the fidelity of an analog radio frequency channel of the FIR filter is improved, and the interference suppression capability is improved.

Description

Technical Field

[0001] The present invention relates to the field of filter optimization design, and more specifically, to a filter multi-dimensional coefficient joint optimization method based on an improved dung beetle algorithm. Background Art

[0002] Since the passive intermodulation (PIM) signals generated by multi-frequency high-power transmitting antennas are easy to pass through the operating frequency of the adjacent receiver, they become interference signals that affect the performance indicators of the receiver. In this case, the receiver can use software to regenerate the transmitted signal to reproduce the passive intermodulation components, align and subtract the interference components from the received signal, and thus improve the receiving performance. In order to minimize the residual interference power after interference subtraction, the receiver can use FIR filters to simulate the RF channel response. At this time, how to determine the optimal coefficients of the FIR filter is the main problem faced.

[0003] At present, the commonly used coefficient optimization method is to collect the passive intermodulation signal at a certain moment, correlate it with the signal generated by filter simulation, and adjust the filter coefficient according to the degree of correlation. However, there are certain differences between the actual collected signal and the simulated signal in practice, and the optimal filter coefficient cannot be obtained based on the above scheme. On the other hand, the filter coefficient can be directly optimized by the ant lion optimization algorithm (ALO), sparrow optimization algorithm (SSA), whale optimization algorithm (WOA), gray wolf optimization algorithm (GWO), etc., through the objective function of maximizing interference suppression performance. However, the above heuristic algorithms are prone to fall into local optimality and it is difficult to obtain the global optimal solution. How to efficiently and accurately jointly optimize the multi-dimensional coefficients of the FIR filter is still a problem to be solved. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a method for joint optimization of multidimensional coefficients of filters based on an improved dung beetle algorithm to solve the problems of poor optimization performance of traditional coefficient optimization methods, uneven global search and local capabilities, easy generation of local optimality, and weak global exploration capability.

[0005] The object of the present invention is achieved through the following solutions:

[0006] A method for joint optimization of multidimensional coefficients of a filter based on an improved dung beetle algorithm comprises the following steps:

[0007] Build an optimization model based on the filter to be optimized to obtain the parameters to be optimized and the optimization objective function;

[0008] Construct a joint optimization method for multidimensional coefficients of filters based on the improved dung beetle algorithm;

[0009] The optimized parameters are introduced into the filter to simulate the generation of passive intermodulation products for interference suppression.

[0010] Furthermore, the filter comprises a FIR filter.

[0011] Furthermore, the optimization model satisfies the following relationship:

[0012] Min AJ(X)

[0013] stlb≤X≤ub

[0014] C(X)=0

[0015] Where AJ(·) is the interference suppression performance, X=[X 1 ,X 2 ,X 3 ,D] is the coefficient of the FIR filter to be optimized, X 1 , X 2 For size N 1 ×1 complex vector, X 3 For size N 2 ×1 real vector symmetric about the center, D is N 3 ×1 vector, lb and ub are the lower and upper limits of each coefficient respectively, and C(X) represents the additional constraints on the coefficient X.

[0016] Furthermore, the improved dung beetle algorithm specifically includes an efficient population initialization strategy based on a suboptimal solution: the filter coefficients are initially optimized based on actual data, and the suboptimal solution and Chebyshev mapping strategy are combined to perform position initialization to reduce the search range and improve the search efficiency. At this time, the initial solution of the i-th dung beetle individual is expressed as:

[0017]

[0018] Among them, Chebyshev k is the Chebyshev chaotic map, expressed as:

[0019] Chebyshev k+1 =cos(4×arccos(Chebyshev k )),Chebyshev k ∈[-1,1];

[0020] is the initial solution of the converted filter coefficients.

[0021] Furthermore, the improved dung beetle algorithm specifically includes a random walk strategy based on the golden sine: the golden sine strategy is used as a random walk strategy during search to improve search speed and search accuracy; the walk strategy based on the golden sine is expressed as:

[0022]

[0023] Among them, δ 1 and δ 2 are the golden ratio coefficients, is the optimal position of the i-th dung beetle individual after t iterations, τ 1 is a random number from (0,2π), τ 2 is a random number in (0,π), ε=rand(1), θ∈[0,π], and when θ=0,π / 2,π, the position remains unchanged; is the current position of the i-th dung beetle individual, is the previous position of the i-th dung beetle individual.

[0024] Furthermore, the improved dung beetle algorithm specifically includes a dynamic weight strategy based on a variant sigmod function: by performing a variant on the sigmod function as a dynamic weight during position update, the algorithm is further used to enhance its ability to coordinate local development and global exploration. The variant sigmod function is expressed as:

[0025]

[0026] Among them, a is the weight amplitude (0,1), b is the weight slope, c is the initial value of the weight, T max is the maximum number of iterations.

[0027] Furthermore, the additional constraint C(X) is directly introduced into the objective function. 1 , X 2 and X 3 The elements in are written as:

[0028]

[0029]

[0030]

[0031] in, and are the real and imaginary parts of the first part of the filter parameters to be optimized respectively; and are respectively the real part and imaginary part of the second part filter parameters to be optimized; N is the third filter parameter to be optimized, and the parameter vector is symmetric about the center; 1 、N 2 and N 3 are the dimensions of the parameters to be optimized respectively;

[0032] At this time, the parameter to be optimized X is expressed as the size after combining the constraint C(X) Real vector of

[0033] Furthermore, the method for jointly optimizing the multidimensional coefficients of the filter based on the improved dung beetle algorithm specifically includes the following sub-steps:

[0034] Use a high-speed oscilloscope to collect passive intermodulation signal samples at a certain moment, and then use the ideal transmission main signal model at that moment as input to simulate the passive intermodulation generation process in the receiver software;

[0035] Compare the actual acquired passive intermodulation signal with the simulated generated signal, and obtain the initial solution of the filter coefficient according to the degree of correlation:

[0036] X init =[X init,1 ,X init,2 ,X init,3 ,D init ]

[0037] According to X 1 , X 2 and X 3 In element form, the filter coefficients are initially solved X init Convert to And combined with Chebyshev chaotic mapping for position initialization.

[0038] Furthermore, when t∈(0,500), take a=0.8, b=0.05, c=0.2.

[0039] The beneficial effects of the present invention include:

[0040] (1) The method of the present invention is based on the dung beetle optimization algorithm, and combines an efficient population initialization strategy based on a suboptimal solution, a random walk strategy based on a golden sine, and a dynamic weight strategy based on a variant sigmoid function to optimize the FIR filter, thereby achieving the best simulation of the simulated RF channel response and improving the interference suppression performance.

[0041] (2) The method of the present invention improves the fidelity of passive intermodulation simulation by improving the coefficients of the FIR filter using the improved dung beetle algorithm. Combined with the collected data, compared with the traditional method, the interference suppression effect of the IDBO algorithm proposed in the method of the embodiment of the present invention is improved by about 1 dB, and compared with other heuristic optimization algorithms (DBO, SSA, ALO, GWO, WOA), the performance is improved by about 0.85 dB at the minimum, and has faster convergence performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0043] Figure 1 FIG. 1 is a diagram of the optimization implementation steps for FIR filter coefficients in an embodiment of the present invention;

[0044] Figure 2 It is a schematic diagram of a multi-dimensional parameter optimization process of a filter based on an improved dung beetle algorithm in an embodiment of the present invention;

[0045] Figure 3 is a variant sigmod function in an embodiment of the present invention;

[0046] Figure 4 Convergence curves of different algorithms;

[0047] Figure 5 Interference suppression performance diagram under different algorithm coefficient optimization in different examples. DETAILED DESCRIPTION

[0048] All features disclosed in all embodiments in this specification, or steps in all methods or processes implicitly disclosed, except for mutually exclusive features and / or steps, can be combined and / or expanded or replaced in any manner.

[0049] The specific implementation process of the present invention is as follows:

[0050] In view of the shortcomings of the existing direct optimization based on correlation degree and traditional coefficient optimization methods, the present invention provides a filter multi-dimensional coefficient joint optimization method based on an improved dung beetle algorithm. Figure 1 As shown, the method comprises the following steps:

[0051] Step 1: construct an optimization model according to the FIR filter to be optimized, and obtain the parameters to be optimized and the optimization objective function;

[0052] Step 2: construct a filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm;

[0053] Step 3: Bring the optimized parameters into the FIR filter to simulate the generation of passive intermodulation components for interference suppression.

[0054] The optimization model satisfies the following relationship:

[0055]

[0056] Where AJ(·) is the interference suppression performance, X=[X 1 ,X 2 ,X 3 ,D] is the coefficient of the FIR filter to be optimized, X 1 , X 2 For size N 1 ×1 complex vector, X 3 For size N 2 ×1 real vector symmetric about the center, D is N 3 × 1 vector. lb and ub are the lower and upper limits of each coefficient, respectively, and C(X) represents an additional constraint on the coefficient X.

[0057] Furthermore, the present invention uses the dung beetle optimization algorithm (DBO) to solve the above problem. The dung beetle algorithm is a new type of swarm intelligence optimization algorithm that optimizes and solves by simulating the rolling, foraging, reproduction and stealing behaviors of dung beetles. At the same time, however, since the coefficients to be optimized in the filter are as many as hundreds of dimensions, it is very easy to fall into the local optimum; in order to improve the global exploration ability and local development ability of the algorithm and avoid the local optimum, the algorithm is improved based on the following strategy.

[0058] Strategy 1: Efficient population initialization strategy based on suboptimal solution: Based on the actual data, the filter coefficients are initially optimized, and the suboptimal solution and Chebyshev mapping strategy are combined to initialize the position, reduce the search range, and improve the search efficiency. At this time, the initial solution of the i-th dung beetle individual is expressed as:

[0059]

[0060] Among them, Chebyshev k is the Chebyshev chaotic map, expressed as:

[0061] Chebyshev k+1 =cos(4×arccos(Chebyshev k )),Chebyshev k ∈[-1,1] (3)

[0062] Strategy 2: Random walk strategy based on golden sine: Using the golden sine strategy as a random walk strategy during search can improve search speed and search accuracy. The walk strategy based on the golden sine is expressed as:

[0063]

[0064] Among them, δ 1 and δ 2 are the golden ratio coefficients, is the optimal position of the i-th dung beetle individual after t iterations. ε=rand(1), θ∈[0,π], and when θ=0,π / 2,π, the position remains unchanged.

[0065] Strategy 3: Dynamic weight strategy based on variant sigmod function: In the early stage of search, we need to improve the exploration efficiency as much as possible to find the optimal solution; in the later stage of search, we need to maintain the exploration rate to avoid local optimality. By making a variant of the sigmod function as a dynamic weight when updating the position, the algorithm's ability to coordinate local development and global exploration is enhanced. The variant sigmod function is expressed as:

[0066]

[0067] The filter multi-dimensional coefficient joint optimization method based on the improved dung beetle algorithm proposed in the present invention can efficiently and accurately search for the optimal coefficients of the FIR filter, improve the fidelity of the FIR filter simulating the radio frequency channel, and improve the interference suppression capability.

[0068] In order to make the purpose, technical method and advantages of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. The specific implementation steps are as follows:

[0069] Step 1: Construct an optimization model based on the FIR filter to be optimized, with X = [X 1 ,X 2 ,X 3 ,D] is the coefficient to be optimized, lb and ub are the lower and upper limits of each coefficient, C(X) represents the additional constraint on coefficient X, and AJ(·) is the weighted sum of interference suppression effects. The optimization model can be expressed as:

[0070]

[0071] Due to X 1 , X 2 For size N 1 ×1 complex vector, X 3 For size N 2 ×1 real vector symmetric about the center, D is N 3 ×1 vector. To facilitate optimization, the additional constraint C(X) is directly introduced into the objective function. 1 , X 2 and X 3 The elements in can be written as:

[0072]

[0073] At this time, the parameter to be optimized X can be expressed as the size after combining the constraint C(X) Real vector of

[0074] Step 2: Use a high-speed oscilloscope to collect samples of the passive intermodulation signal at a certain moment, and then use the ideal transmission main signal model at that moment as input to simulate the passive intermodulation generation process in the receiver software. Compare the actual acquired passive intermodulation signal with the simulated generated signal, and obtain the initial solution of the filter coefficient according to the degree of correlation:

[0075] X init =

X init,1 ,X init,2 ,X init,3 ,D init

[0076] According to formula (7), the initial solution of filter coefficients X init Convert to Combined with Chebyshev chaotic mapping, the position is initialized by equation (2).

[0077] Step 3: In order to quickly search for the optimal solution in the feasible solution space, the dung beetle will randomly walk in the feasible solution. In order to improve the search capability, the present invention introduces the golden sine strategy as the walking strategy of the dung beetle, as shown in formula (4).

[0078] Step 4: In order to ensure the efficiency of each search, the "breeding" area can be sought based on the boundary selection strategy to ensure that the new individual is born in a better area. After the new individual is born, it is necessary to find the best foraging area to guide it to forage, that is, to find the area with the optimal solution.

[0079] Step 5: In order to efficiently search near the current optimal solution, some "thieving dung beetles" can be used to "compete" for positions near the optimal solution. At the same time, in order to ensure the efficiency of the search, the exploration efficiency needs to be improved as much as possible in the early stage of the search to find the optimal solution; and in the later stage of the search, the exploration rate needs to be maintained to avoid falling into the local optimum. Therefore, a dynamic weight strategy based on a variant sigmoid function is considered as the position update strategy of the "thieving dung beetle", as shown in formula (5). When t∈(0,500), take a=0.8, b=0.05, c=0.2, sigmoid 1 (t) and sigmoid 2 (t) Figure 3 shown.

[0080] At this time, the position update of the "thieving dung beetle" can be expressed as:

[0081]

[0082] Among them, b 3 For size An independent random vector of .

[0083] Step 6: According to step 1-step 5, first initialize the dung beetle population and the number of algorithm adoptions; calculate the objective function values ​​of all dung beetles according to the objective function; update the positions of all dung beetles and determine whether they are outside the boundary; update the current optimal solution and the corresponding objective function value, and repeat the above steps until the termination condition is met, and output the global optimal solution According to formula (7), the filter coefficients are restored and the optimal filter coefficients X are obtained. opt The whole algorithm process is as follows Figure 2 shown.

[0084] The performance results of the filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm in the example of the present invention are as follows: Figure 4 , Figure 5 As shown in Table 1. From the simulation results, it can be seen that compared with the traditional method, the interference suppression effect of the IDBO algorithm (self-named term) proposed in the present invention is improved by about 1 dB, and compared with other heuristic optimization algorithms (DBO, SSA, ALO, GWO, WOA), the performance is improved by about 0.85 dB at the lowest, and the algorithm converges faster.

[0085] Table 1 Interference suppression performance under filter coefficient optimization of different algorithms

[0086]

[0087] The units involved in the embodiments of the present invention may be implemented by software or hardware, and the units described may also be arranged in a processor. The names of these units do not, in some cases, limit the units themselves.

[0088] According to one aspect of an embodiment of the present invention, a computer program product or a computer program is provided, the computer program product or the computer program includes a computer instruction, and the computer instruction is stored in a computer-readable storage medium. A processor of a computer device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the method provided in the above various optional implementations.

[0089] As another aspect, an embodiment of the present invention further provides a computer-readable medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device implements the method described in the above embodiment.

Claims

1. A filter multidimensional coefficient joint optimization method based on an improved dung beetle algorithm, characterized in that: The following steps are involved: Build an optimization model based on the filter to be optimized to obtain the parameters to be optimized and the optimization objective function; Construct a joint optimization method for multidimensional coefficients of filters based on the improved dung beetle algorithm; The optimized parameters are introduced into the filter to simulate the generation of passive intermodulation products for interference suppression.

2. The filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm according to claim 1 is characterized in that: The filter comprises a FIR filter.

3. The filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm according to claim 2 is characterized in that: The optimization model satisfies the following relationship: Min AJ(X) stlb≤X≤ub C(X)=0 Wherein, AJ(·) is the interference suppression performance, X = [X1, X2, X3, D] is the coefficient to be optimized of the FIR filter, X1, X2 are complex vectors of size N1×1, X3 is a real vector of size N2×1 symmetric about the center, D is a vector of size N3×1, lb and ub are the lower and upper limits of each coefficient, respectively, and C(X) represents the additional constraints on the coefficient X.

4. The filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm according to claim 3 is characterized in that: The improved dung beetle algorithm specifically includes an efficient population initialization strategy based on a suboptimal solution: a filter coefficient is initially optimized based on actual data, and a suboptimal solution and a Chebyshev mapping strategy are combined to initialize the position, so as to reduce the search range and improve the search efficiency. At this time, the initial solution of the i-th dung beetle individual is expressed as: Among them, Chebyshev k is the Chebyshev chaotic map, expressed as: Chebyshev k+1 =cos(4×arccos(Chebyshev k )),Chebyshev k ∈[-1,1]; is the initial solution of the converted filter coefficients.

5. The filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm according to claim 4 is characterized in that: The improved dung beetle algorithm specifically includes a random walk strategy based on the golden sine: the golden sine strategy is used as a random walk strategy during search to improve search speed and search accuracy; the walk strategy based on the golden sine is expressed as: Among them, δ1 and δ2 are the golden section coefficients, is the optimal position of the i-th dung beetle individual after t iterations, τ1 is a random number in (0,2π), τ2 is a random number in (0,π), ε=rand(1), θ∈[0,π], and when θ=0,π / 2,π, the position remains unchanged; is the current position of the i-th dung beetle individual, is the previous position of the i-th dung beetle individual.

6. The filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm according to claim 1 is characterized in that: The improved dung beetle algorithm specifically includes a dynamic weight strategy based on a variant sigmod function: by performing a variant on the sigmod function as a dynamic weight during position update, the algorithm is further used to enhance its ability to coordinate local development and global exploration. The variant sigmod function is expressed as: Among them, a is the weight amplitude (0,1), b is the weight slope, c is the initial value of the weight, T max is the maximum number of iterations.

7. The filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm according to claim 4 is characterized in that: The additional constraint C(X) is directly introduced into the objective function, and the elements in X1, X2 and X3 are written as: in, and are the real and imaginary parts of the first part of the filter parameters to be optimized respectively; and are respectively the real part and imaginary part of the second part filter parameters to be optimized; is the third part of the filter parameters to be optimized, and the parameter vector is symmetric about the center; N1, N2 and N3 are the dimensions of the corresponding parameters to be optimized; At this time, the parameter to be optimized X is expressed as the size after combining the constraint C(X) The real vector of 8. The filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm according to claim 7, characterized in that: The method for jointly optimizing the multidimensional coefficients of the filter based on the improved dung beetle algorithm specifically includes the following sub-steps: Use a high-speed oscilloscope to collect passive intermodulation signal samples at a certain moment, and then use the ideal transmission main signal model at that moment as input to simulate the passive intermodulation generation process in the receiver software; Compare the actual acquired passive intermodulation signal with the simulated generated signal, and obtain the initial solution of the filter coefficient according to the degree of correlation: X init =[X init,1 ,X init,2 ,X init,3 ,D init ] The filter coefficients are initially solved X according to the element form in X1, X2 and X3. init Convert to And combined with Chebyshev chaotic mapping for position initialization.

9. The filter multidimensional coefficient joint optimization method based on the improved dung beetle algorithm according to claim 6, characterized in that: When t∈(0,500), take a=0.8, b=0.05, c=0.2.

Images