CN-122017827-A - Micro Doppler frequency suppression method based on adaptive successive variation modal decomposition

CN122017827ACN 122017827 ACN122017827 ACN 122017827ACN-122017827-A

Abstract

The invention discloses a micro Doppler frequency suppression method based on adaptive Successive Variation Modal Decomposition (SVMD). The method takes spectrum entropy minimization as an optimization target, utilizes Particle Swarm Optimization (PSO) to adaptively search and determine SVMD optimal cost parameters, and simultaneously realizes adaptive determination of the number of the decomposition modal functions. Under the condition of optimal parameters, carrying out modal decomposition on radar echo signals, taking the modal component with the largest energy as a main body signal according to an energy judgment criterion, and inhibiting the rest modal components as micro-motion signals, thereby realizing the effective separation of micro-Doppler components and the main body signal. The simulation data and the actually measured radar data are processed and analyzed, and the result shows that the method can effectively inhibit micro Doppler frequency, improve the definition and the resolvable property of the main body signal, and has higher practical value.

Inventors

LI WEIDONG
WANG RUI
HU CHENG
Gong Chunxu

Assignees

北京理工大学长三角研究院(嘉兴)
北京理工大学

Dates

Publication Date: 20260512
Application Date: 20260112

Claims (5)

1. The micro Doppler frequency suppression method based on the adaptive successive variation modal decomposition is characterized by comprising the following steps of: S1, acquiring complex radar echo data, respectively carrying out projection operation along a plurality of preset different projection directions, and extracting the real part of each projection direction signal as an input signal of subsequent decomposition; s2, setting the number of initial decomposition modes as 1, and respectively processing each projection real part signal acquired in the S1; S3, extracting a modal component with highest energy by utilizing SVMD algorithm under the optimal parameter condition of S2 as a main signal, judging whether the currently obtained modal component meets the requirement of signal separation according to a preset threshold judgment condition, if so, entering a step S4, if not, adding 1 to the number of the modalities, and returning to the step S2 to continue decomposition until the judgment condition is met or the maximum decomposition modal number is reached; and S4, combining all modal components obtained in each projection direction, selecting the modal component with the highest energy value as a main body signal according to an energy judgment criterion, and inhibiting the rest modal components as micro Doppler signals, thereby completing the effective separation of the micro signals and the main body signal.
2. The micro-doppler frequency suppression method based on adaptive successive variation modal decomposition according to claim 1, wherein in the step S1, the signal is projected to the mth direction to obtain a real value signal: (1); Wherein, the Is a signal In the direction of As an input signal for subsequent decomposition.
3. The micro-doppler frequency suppression method based on adaptive successive variation modal decomposition according to claim 1, wherein the step S2 specifically includes: (1) Setting basic parameters of PSO including cost parameter search range Maximum iteration number T, particle swarm size N, etc.; (2) For the ith particle ) Corresponding cost parameter Performing SVMD decomposition to obtain K eigenmode components ; (2); Taking the highest energy modal component Is of the frequency spectrum of (a) As a main signal spectrum, calculating an entropy value H as an fitness value of the ith particle by the formula (3); (3); Finding out the position with the minimum fitness value as the global optimal position, namely the current optimal cost parameter; (3) Repeating the step (2) by updating the position of the particles, namely the value of the cost parameter, comparing the change of the fitness value, and continuously adjusting the global optimal position and the particle updating direction (4); Wherein, the A cost parameter representing particle i in the t-th iteration, Indicating the speed of the vehicle and, A historical optimal position for the device; the global optimal position in the t-th iteration; 、 is a random number between [0, 1]; 、 is a learning factor for regulating the degree of dependence of particles on individual and population experience; is inertial weight and used for balancing global searching and local searching capacity, and is larger Is favorable for global searching, jumps out of local extremum, does not fall into local optimum, and is smaller The method is favorable for local search, so that the algorithm can be quickly converged to an optimal solution, when the actual optimization problem is solved, the method is usually hoped to quickly converge the search space to a certain area by adopting global search and then obtain a high-precision solution by adopting local fine search, so that an adaptive adjustment strategy is provided, namely the method linearly reduces as iteration progresses The iterative formula is formula (5): (5); Wherein, the And Respectively representing a maximum inertial weight and a minimum inertial weight; Representing the number of current iterations and, Representing a maximum number of iterations; (4) And stopping iteration when the iteration number reaches the set upper limit T, and outputting a cost parameter corresponding to the global optimal particle position, wherein the cost parameter is the optimal cost parameter decomposed by SVMD.
4. The micro-doppler frequency suppression method based on adaptive successive variation modal decomposition according to claim 1, wherein in the step S3, the concrete flow of SVMD algorithm is as follows: (1) Taking a real value signal obtained by projection in the mth direction obtained in the S1 As a signal to be decomposed, the optimal particle position obtained in S2 is taken as a cost parameter The initialization parameters comprise the decomposed modal number L=1, the iteration number n=0 and the initial center frequency Preset threshold value 、 Lagrangian multiplier Etc.; (2) Updating the decomposed L-th modality function by equation (6) ; (6); Wherein, the For the modal function obtained for the n+1th iteration, Is that Is a representation of the frequency domain of (a), The center frequency of the L-th modal function for the nth iteration, Is the Lagrangian multiplier after the nth iteration; After modal update, new center frequency is calculated through weighting of spectrum energy distribution : (7); And updates the Lagrangian multiplier accordingly : (8); Wherein, the Continuously increasing the value of n, and circularly executing the step (2) until the difference between adjacent iteration modes meets the formula (9): (9); (3) After each mode is extracted, judging whether the total reconstruction error of the current mode decomposition meets a set precision threshold or not, if yes, satisfying a formula (10): (10); Wherein, the If the condition is met, the whole decomposition process is terminated, otherwise, the next mode is continuously extracted until the termination criterion is met; The above is to real value signal Performing SVMD decomposition to obtain L modal component sets ; (11)
5. The adaptive successive variation modal decomposition-based micro-doppler frequency suppression method as claimed in claim 1, wherein in step S4, the complex signals are decomposed to form a complex set Expressed as: (12); at the collection The mode component with the highest energy is found to be the needed main body signal, so that the micro Doppler frequency separation is realized.

Description

Micro Doppler frequency suppression method based on adaptive successive variation modal decomposition Technical Field The invention belongs to the technical field of radar signal processing, and particularly relates to a micro Doppler frequency suppression method based on adaptive successive variation modal decomposition. Background In Inverse Synthetic Aperture Radar (ISAR) imaging, complex objects often contain locally moving parts whose echo signals reflect not only the rigid motion of the object as a whole, but also the disturbance component caused by the locally non-rigid motion. Such as rotation of the rotor of the drone, rotation of the vehicle tires, and swing of the limbs of the human body during walking or running, etc. Such local motion is collectively referred to as "micro-motion", which introduces additional doppler frequency components in the echo signal, a phenomenon defined as the "micro-doppler effect". In typical ISAR imaging, if the rotation angle of the target in the coherent accumulation time is small, the target can be regarded as a rigid body, and the azimuth information of the target can be acquired through doppler modulation of the target, so that high-resolution imaging is realized. However, for targets with significant jogs, the target no longer satisfies the rigid body assumption, and if the target is still processed by using the conventional ISAR imaging algorithm, the frequency variation introduced by the jog component will form an interference band extending along the azimuth direction in the imaging result, so that the imaging quality of the target main body is seriously affected. Therefore, there is a need to effectively separate the main component from the micro-motion component in the radar echo to eliminate the micro-motion interference and improve the imaging quality. However, the existing methods still have limitations. The decomposition performance of the system is highly dependent on two key parameters, namely the modal decomposition number and cost parameter. These parameters are typically set manually, empirically, or obtained by trial and error, and lack global optimization capabilities and adaptive mechanisms. Once the parameters are improperly selected, the problems of modal leakage, excessive decomposition or insufficient decomposition are easily caused, so that the accurate extraction of micro-Doppler components and the separation quality of the whole signals are affected, and a self-adaptive method for inhibiting the micro-Doppler frequency is needed. Disclosure of Invention In order to solve the problems, the invention provides a micro Doppler frequency suppression method based on adaptive successive variation modal decomposition. According to the method, spectrum entropy is used as an objective function, a PSO algorithm is introduced to perform self-adaptive search and optimization on cost parameters in SVMD, self-adaptive determination of the number of decomposition mode functions is realized on the basis, and finally effective separation of a inching signal and a main body signal is realized without manually setting parameters. The technical scheme of the invention is as follows: The micro Doppler frequency suppression method based on the adaptive successive variation modal decomposition comprises the following steps: s1 for received radar echo signals It is necessary to convert to real signal processing first. An initial projection parameter M (i.e., the number of projections) is set, and the value of M is set from 1 to M. Projecting the signal in the mth direction to obtain a real value signal: (1) Wherein, the Is a signalIn the direction ofAs an input signal for subsequent decomposition. S2, for projection signalsAnd (5) performing a PSO optimization algorithm to determine an optimal cost parameter. (1) Setting basic parameters of PSO including cost parameter search rangeThe maximum iteration number T, the particle swarm size N, etc. (2) For the ith particle) Corresponding cost parameterPerforming SVMD decomposition to obtain K eigenmode components。 (2) Taking the highest energy modal componentIs of the frequency spectrum of (a)As a main signal spectrum, an entropy value H is calculated by the formula (3) as an fitness value of the i-th particle. (3) And finding the position with the minimum fitness value as the global optimal position, namely the current optimal cost parameter. (3) Repeating the step (2) by updating the position of the particles, namely the value of the cost parameter, comparing the change of the fitness value, and continuously adjusting the global optimal position and the particle updating direction (4) Wherein, the A cost parameter representing particle i in the t-th iteration,Indicating the speed of the vehicle and,A historical optimal position for the device; the global optimal position in the t-th iteration; 、 is a random number between [0, 1]; 、 is a learning factor for regulating the degree of dependence of particles on individ