CN-121541167-B - Multi-target detection method and device for pulse Doppler radar in complex electromagnetic interference environment

CN121541167BCN 121541167 BCN121541167 BCN 121541167BCN-121541167-B

Abstract

The invention discloses a multi-target detection method in a complex electromagnetic interference environment aiming at a pulse Doppler radar, which is characterized in that a joint back diffusion process is introduced to update an interference sample and an amplitude sample in a joint way, so that the problems that a target signal is covered or submerged by an interference signal, the detection false alarm rate rises, interference and a target are difficult to effectively separate are solved, a neural network of an interference score function is trained by a denoising score matching criterion is introduced to solve the problems of high interference intensity and complex structure in the multi-target detection in the complex electromagnetic interference environment, and a sparse Bayesian learning method is introduced to adaptively model the sparse structure of a multi-target echo signal. The invention also provides a multi-target detection device under the complex electromagnetic interference environment. The method provided by the invention can realize modeling and suppression of multiple types of interference, and simultaneously improves detection probability, interference suppression capability and system robustness under complex electromagnetic environment on a plurality of low observable targets in distance and Doppler dimensions.

Inventors

ZHU JIANG
GUO RUOHAI
ZHANG NING
WANG ZHIGANG
WANG XINHAI
QU FENGZHONG

Assignees

浙江大学
中国船舶集团有限公司第七二四研究所

Dates

Publication Date: 20260508
Application Date: 20260116

Claims (6)

1. The multi-target detection method for the pulse Doppler radar in the complex electromagnetic interference environment is characterized by comprising the following steps of: step 1, obtaining interference sample data to construct a fractional function neural model for generating interference variables; step 2, acquiring echo signals and constructing complex amplitude vectors corresponding to the echo signals by matching with an overcomplete dictionary matrix, wherein the construction process of the complex amplitude vectors is as follows: Digital wave beam forming is carried out on the echo signals so as to obtain detection information in a fast time dimension and a slow time dimension; constructing complex amplitude vectors corresponding to echo signals based on the detection information and the overcomplete dictionary matrix; the detection information is obtained through two-dimensional data formed by fast time and slow time of a straightening target; Step3, constructing a joint back diffusion process, and initializing a complex amplitude vector and interference sample data as an initial sampling starting point, wherein the joint back diffusion process is described by adopting a random differential equation, and the expression is as follows: wherein, the method comprises the steps of, The gradient is represented by a gradient, Is a joint posterior distribution of amplitude and interference, Is a standard wiener process in the reverse direction, Is that The individual targets correspond to complex amplitude vectors of different doppler-range meshes; the expression of the initial sampling start point is as follows: , wherein, the method comprises the steps of, Represent the first A number of complex amplitude vector samples, Represent the first The number of interference samples is chosen to be less than the number of interference samples, Is a transition probability parameter at time 1, Representing dimensions as Is used for the matrix of units of (a), Representing dimensions as Step 4, taking time as a node, executing an corrector, namely respectively updating the fractional functions of the complex amplitude vector and the interference variable aiming at the sample at the current moment, and correcting the updated sample by adopting Lang-N dynamics criterion to obtain a sample after sampling and correction; step 5, updating the prior variance of the complex amplitude vector in the step 4 by using sparse Bayesian learning, and updating the fraction of the complex amplitude vector; step 6, executing a predictor: Calculating the gradient of the joint posterior distribution on the complex amplitude vector and the disturbance variable based on the fraction obtained in the step 5, and generating a reconstructed sample according to a back diffusion random differential equation by utilizing the gradient; Updating prior variances of complex amplitudes in the reconstructed samples by using sparse Bayesian learning; step 7, repeating the step 4-6 until the diffusion termination time to obtain a complex amplitude estimated value corresponding to the sample; and 8, inputting the complex amplitude estimation value to a constant false alarm detector to construct an effective target set.
2. The method for multi-target detection in a complex electromagnetic interference environment for pulse doppler radar according to claim 1, wherein the fractional function neural model is obtained by constructing a diffusion model and a denoising fractional matching criterion.
3. The method for multi-target detection in a complex electromagnetic interference environment for pulsed doppler radar according to claim 2, wherein the expression of the denoising score matching criterion is as follows: wherein, the method comprises the steps of, As a parameter of the neural network, It is indicated that the desire is to be met, Representation of Is in the interval Is obtained by uniformly sampling the materials in the middle, Is sampled from the set of interference samples, Representing the L2 norm.
4. The method for multi-target detection in a complex electromagnetic interference environment for pulsed doppler radar of claim 1, wherein the overcomplete dictionary matrix is constructed based on a doppler network and a range network.
5. The method for multi-target detection in a complex electromagnetic interference environment for pulsed doppler radar according to claim 1, wherein the complex amplitude estimation value is obtained by calculating an average value of all complex amplitude vectors output at a diffusion termination time.
6. A multi-target detection device in a complex electromagnetic interference environment, characterized by performing the steps of the multi-target detection method in a complex electromagnetic interference environment for a pulse doppler radar according to any one of claims 1 to 5.

Description

Multi-target detection method and device for pulse Doppler radar in complex electromagnetic interference environment Technical Field The invention belongs to the technical field of radar signal processing, and particularly relates to a multi-target detection method and device in a complex electromagnetic interference environment aiming at a pulse Doppler radar. Background The pulse Doppler radar is used as an important system for high-resolution and high-sensitivity target detection, and has wide prospects in the applications of air monitoring, low-altitude defense, unmanned aerial vehicle countering and the like. However, in practical application scenarios, the radar faces serious challenges of complex electromagnetic interference environments, including structured interference signals from sources such as communication systems, power equipment, electronic countermeasure platforms, etc., such as communication, frequency modulation, phase modulation, amplitude modulation, comb spectrum interference, etc. The interference signals have strong energy, so that the target echo signals are seriously submerged or covered, the radar detection performance is obviously reduced, and the problems of target missed detection, tracking loss, great increase of false alarm rate and the like are presented. When interference exists, the traditional linear signal processing method, namely pulse compression and moving target detection, cannot fully distinguish characteristics of targets and interference based on a matched filtering mode and a frequency domain filtering mode, and accurate modeling and suppression of non-stationary or non-Gaussian interference are difficult. In addition, most of these conventional methods rely on a priori knowledge of the interference characteristics, which are difficult to adapt to complex electromagnetic interference environments, resulting in serious degradation of multi-target detection performance. Patent document CN120294733a discloses a radar one-dimensional range profile multi-target detection method based on CFAR and isolated forest fusion, which comprises the steps of obtaining radar raw data and preprocessing the radar one-dimensional range profile data, detecting targets in the radar one-dimensional range profile data by utilizing CACFAR detection self-adaptive threshold technology and normalizing energy values, calculating target scores by an isolated forest method, weighting and fusing the CACFAR detection and isolated forest results to obtain fusion results, defining a new detection threshold to confirm targets, setting a new detection threshold and judging whether targets exist or not. Patent document CN119291641a discloses a multi-target detection method and system based on millimeter wave radar, which comprises the steps of acquiring original echo data acquired by the millimeter wave radar, wherein the original echo data comprises target data points of a plurality of targets, selecting a preset number of data points from the target data points of the targets, taking the selected data points as effective data points, and detecting the targets based on the effective data points. Disclosure of Invention The invention aims to provide a multi-target detection method and device in a complex electromagnetic interference environment aiming at a pulse Doppler radar, which can realize modeling and suppression of multiple types of interference (such as comb spectrum interference, continuous wave interference, linear frequency modulation interference and the like) and simultaneously improve detection probability, interference suppression capability and system robustness of a plurality of low observable targets in distance and Doppler dimensions in the complex electromagnetic environment. In order to achieve the first object of the present invention, a multi-target detection method in a complex electromagnetic interference environment for a pulse doppler radar is provided, comprising the steps of: Step 1, obtaining interference sample data, and constructing a score function neural model for generating interference variables based on a diffusion model and a denoising score matching criterion; Step 2, acquiring echo signals transmitted by a pulse Doppler radar in a complex electromagnetic interference environment, and performing digital beam forming on the echo signals to obtain detection information in a fast time dimension and a slow time dimension; constructing complex amplitude vectors corresponding to echo signals based on the detection information and the overcomplete dictionary matrix; step 3, constructing a joint back diffusion process, and initializing complex amplitude vectors and interference sample data as initial sampling starting points; And 4, taking time as a node, and executing the corrector: updating the fraction function of the complex amplitude vector and the fraction function of the disturbance variable aiming at the sample at the current moment, and correcting the updated