阅读说明：本技术 基于稀疏表示的无线电引信抗箔条干扰方法 (Radio fuse foil strip interference resisting method based on sparse representation ) 是由 刘景萍 张帆 林慧君 刘遨云 常梦璐 李紫婷 李文东 秦港 于 2019-08-31 设计创作，主要内容包括：本发明公开一种基于稀疏表示的无线电引信抗箔条干扰方法,包括如下步骤：(10)预处理：对引信回波信号进行归一化处理,选取目标回波信号和箔条干扰信号得到两种训练样本；(20)字典学习：基于K聚类奇异值分解方法对两种训练样本分别学习,得到两种自适应稀疏字典；(30)稀疏表示：基于正交匹配跟踪算法,输出稀疏系数；(40)阈值判决：根据预设阈值,判断引信回波信号类型；(50)结果输出：若是引信目标回波,则直接输出该重构信号,反之,则将回波信号与重构信号作差,得到引信目标回波。本发明的无线电引信抗箔条干扰方法,自适应能力较强、实时性好,处理速度较快。(The invention discloses a radio fuse foil strip interference resisting method based on sparse representation, which comprises the following steps of: (10) pretreatment: normalization processing is carried out on the fuze echo signal, and a target echo signal and a foil strip interference signal are selected to obtain two training samples; (20) dictionary learning: respectively learning the two training samples based on a K clustering singular value decomposition method to obtain two self-adaptive sparse dictionaries; (30) sparse representation: outputting a sparse coefficient based on an orthogonal matching tracking algorithm; (40) and (3) threshold judgment: judging the type of a fuse echo signal according to a preset threshold value; (50) and (4) outputting a result: and if the echo signal is the fuze target echo, directly outputting the reconstruction signal, otherwise, subtracting the echo signal from the reconstruction signal to obtain the fuze target echo. The method for resisting the interference of the foil strips by the radio fuse has the advantages of strong self-adaptive capacity, good real-time performance and high processing speed.)

1. A radio fuse foil strip interference resisting method based on sparse representation is characterized by comprising the following steps:

(10) pretreatment: normalization processing is carried out on the fuse echo signals to obtain sparsely decomposed fuse echo signals, and two training samples are obtained by selecting target echo signals and foil strip interference signals;

(20) dictionary learning: respectively learning the two training samples based on a K clustering singular value decomposition method to obtain two self-adaptive sparse dictionaries: a target sparse dictionary and a foil strip interference sparse dictionary;

(30) sparse representation: based on an orthogonal matching tracking algorithm, performing sparse representation on the fuze echo signal after the normalization processing by using a target sparse dictionary, and outputting a sparse coefficient; or sparse representation is carried out on the fuze echo signals after normalization processing by using a foil sparse dictionary, and a sparse coefficient is output;

(40) and (3) threshold judgment: vector-based l₂Calculating to obtain a sparse reconstruction error by using a fuse echo signal, a sparse coefficient and a sparse dictionary according to a norm theory, and judging the type of the fuse echo signal according to a preset threshold value;

(50) and (4) outputting a result: and if the echo signal is the fuze target echo, directly outputting the reconstruction signal, otherwise, subtracting the echo signal from the reconstruction signal to obtain the fuze target echo.

2. The radio fuze anti-foil strip interference method according to claim 1, characterized in that the step of (20) sparse dictionary learning comprises:

(21) pretreatment: inputting a training set and a blank dictionary, initializing the blank dictionary and the cycle number, and setting an upper limit of the cycle number and a reconstruction error threshold;

(22) sparse coding: obtaining a sparse coefficient of a training sample by using an orthogonal matching tracking algorithm;

(23) and (3) dictionary updating: updating each atom in the dictionary and the corresponding sparse coefficient: firstly, finding out corresponding index set numbers of all training samples using atoms in a dictionary; calculating by utilizing atoms and sparse representation coefficients to obtain an error matrix; finding out the index number corresponding to the error matrix; finally, singular value decomposition is carried out to update dictionary atoms and sparse coefficients;

(24) and (3) loop iteration: after one-time sparse coding and dictionary updating are completed, the cycle times are increased automatically, whether an iteration condition is met or not is calculated, the iteration is stopped if the iteration condition is met, and otherwise, the execution is skipped (22);

(25) and (3) dictionary output: and obtaining and outputting a coefficient dictionary.

3. The radio fuze anti-foil interference method according to claim 1, wherein in the (30) sparse representation step, the sparse representation with the target sparse dictionary comprises:

(31) pretreatment: inputting a target sparse dictionary, a fuse echo signal to be detected and sparsity, and initializing cycle times, sparse coefficients, an index set and residual initial values;

(32) updating an index set: calculating the maximum component index by using the sparse dictionary and the residual error, and completing updating;

(33) residual error updating: calculating an index column in the sparse dictionary and a signal after sparse coefficient reconstruction, and performing difference on a fuse echo signal to be detected to obtain a residual error;

(34) and (3) loop iteration: performing self-increment operation of the cycle times, calculating whether the sparsity condition is met, if so, stopping iteration, otherwise, skipping execution (32);

(35) and (3) coefficient output: and obtaining and outputting a sparse coefficient.

Technical Field

The method belongs to the technical field of radio fuze interference, and particularly relates to a sparse representation-based radio fuze foil strip interference resisting method which is strong in self-adaptive capacity, good in real-time performance and high in processing speed.

Background

Foil strips are often put in electronic warfare to resist interference, the foil strips are an electronic interference measure with simple technology, low cost and obvious effect, and the action mechanism is that a large number of foil strips which are randomly distributed are put in the air to form a foil strip cloud or an interference corridor so as to confuse, shield or weaken local electronic weapons and achieve the effect of protecting own warfare tools and personnel safety. The position of foil strip interference is more and more important since world war II, and due to the remarkable interference performance, the research on foil strip interference technology and anti-interference method is competitively developed in various countries and regions.

The radio fuse is a device for detecting a target by utilizing the electromagnetic scattering property of the target and controlling the detonation of ammunition, is one of the important components of a weapon system, and is widely applied to various military fields of aerospace, aviation, chemical defense and the like and civil occasions of automobile collision prevention, speed measurement and distance measurement and the like. In order to solve the problems that the radio fuse is interfered by a foil strip when detecting a target, the target identification is wrong, and an enemy target cannot be accurately and effectively attacked, a radio fuse foil strip interference resisting technology must be deeply researched, so that the radio fuse foil strip interference resisting technology has profound significance for improving the performance of a weapon system and consolidating national defense safety to a certain extent.

The research on foil strip interference resistance, the existing methods and technologies in the literature and the patent at present mainly comprise: (1) a time-frequency domain filtering technique; (2) a polarization recognition method; (3) doppler filtering; (4) a signal feature identification method; (5) improving radar or fuze resolution; (6) composite guidance techniques, and the like. For example, in the document "application of time-frequency analysis technology in suppressing foil interference", a time-frequency domain filtering technology is adopted, and a linear and time-varying filter is designed in a specific time-frequency pass domain and is used for filtering in a time-frequency joint domain. By taking reference to the time-frequency projection and time-frequency expansion theory, the method is provided for designing a domain-pass and time-frequency projection filter of a target range profile based on Wigner-Ville distribution of a linear signal space and performing domain-pass filtering in a pulse period on radar echoes, and has the capability of inhibiting interference under the background of foil clutter; for another example, the patent "pulse doppler radar polarization anti-interference method" adopts a polarization identification method, and the technical scheme is as follows: inputting a radar dual-polarization echo signal, judging whether active suppression interference exists or not, and performing polarization cancellation when the suppression interference exists; then calculating the moving time of the target signal between the front pulse and the rear pulse, and carrying out distance moving correction on the pulse sequence; then performing coherent accumulation on the dual-polarized channel, shifting a zero frequency to the center of a frequency spectrum, locking a channel number of a target in a Doppler-distance two-dimensional matrix, setting a target two-dimensional tracking wave gate, and performing two-dimensional constant false alarm detection in the tracking wave gate; and finally, carrying out target polarization ratio detection judgment, and outputting a signal passing through the detection judgment as a target signal after interference suppression, thereby improving the correct detection probability of the target.

The problems existing in the prior method and the prior art are mainly as follows: (1) the method can not adapt to and meet the actual combat environment: most methods only aim at a certain specific scene, verify that the simulation data comes from, and have great difference with the actual combat environment; (2) the real-time performance is poor: the system requirements cannot be met, and the algorithm performance needs to be improved; (3) the engineering property is poor: hardware realization and test verification links are omitted, and the method mainly stays in an algorithm theory research stage; (4) at present, most research objects are radars, and relatively few studies on foil strip interference resistance of radio fuses are carried out.

Disclosure of Invention

The invention aims to provide a radio fuze foil strip interference resisting method based on sparse representation, which has the advantages of strong self-adaptive capacity, good real-time performance and high processing speed.

The technical solution for realizing the purpose of the invention is as follows:

a radio fuze foil strip interference resisting method based on sparse representation comprises the following steps:

(10) pretreatment: normalization processing is carried out on the fuse echo signals to obtain sparsely decomposed fuse echo signals, and two training samples are obtained by selecting target echo signals and foil strip interference signals;

(20) sparse dictionary learning: respectively learning the two training samples based on a K clustering singular value decomposition method to obtain two self-adaptive sparse dictionaries: a target sparse dictionary and a foil strip interference sparse dictionary;

(30) sparse representation: based on an orthogonal matching tracking algorithm, performing sparse representation on the fuze echo signal after the normalization processing by using a target sparse dictionary, and outputting a sparse coefficient; or sparse representation is carried out on the fuze echo signals after normalization processing by using a foil sparse dictionary, and a sparse coefficient is output;

(40) and (3) threshold judgment: vector-based l₂Calculating to obtain a sparse reconstruction error by using a fuse echo signal, a sparse coefficient and a sparse dictionary according to a norm theory, and judging the type of the fuse echo signal according to a preset threshold value;

(50) and (4) outputting a result: and if the echo signal is the fuze target echo, directly outputting the reconstruction signal, otherwise, subtracting the echo signal from the reconstruction signal to obtain the fuze target echo.

Compared with the prior art, the invention has the following remarkable advantages:

1. the self-adaptive capacity is stronger, can satisfy the requirement of complicated combat environment: the sparse dictionary is obtained by learning a training sample constructed on the basis of data obtained by testing in a real scene, has stronger sparse representation capability on echo signals mixed with interference signals, and can meet the anti-interference treatment of different foil strip interference scenes and real combat environments as long as an over-complete dictionary library is abundant enough.

2. The real-time property is good, and the processing speed is fast: the method for establishing the overcomplete dictionary base by using the offline learning dictionary does not occupy the anti-interference signal processing time, and meanwhile, the orthogonal matching tracking algorithm has higher convergence speed, fewer atoms to be reconstructed and superior algorithm performance;

3. the feasibility of engineering realization is higher: the method mainly focuses on training sample collection, learning and storage of the sparse dictionary and a signal processing module, and a hardware implementation scheme is practical and effective, and has strong testability and higher engineering implementation feasibility.

4. The method is favorable for perfecting the radio fuse anti-interference system: the radar and the fuse have certain common points in the aspects of working principle and signal analysis, but because the fuse is terminal guidance, the acting distance and the acting time are short, and the requirements on the aspects of instantaneity, identification precision, interference resistance and the like are more strict than those of a radar system. The research object of the invention changes from radar to fuze, and the deep research on the anti-interference technology of the radio fuze can improve the fighting efficiency of the fuze to a certain extent and enrich the electronic countermeasure knowledge system.

Drawings

Fig. 1 is a main flow chart of a radio fuse foil strip interference resisting method based on sparse representation.

Fig. 2 is a flowchart of the sparse dictionary learning step of fig. 1.

Fig. 3 is a flow chart of the sparse representation step of fig. 1.

Detailed Description

As shown in fig. 1, the radio fuze foil strip interference resisting method based on sparse representation of the present invention includes the following steps:

(10) pretreatment: normalization processing is carried out on the fuse echo signals to obtain sparsely decomposed fuse echo signals, and two training samples are obtained by selecting target echo signals and foil strip interference signals;

(20) sparse dictionary learning: respectively learning the two training samples based on a K clustering singular value decomposition method to obtain two self-adaptive sparse dictionaries: a target sparse dictionary and a foil strip interference sparse dictionary;

as shown in fig. 2, the (20) sparse dictionary learning step includes:

(21) pretreatment: inputting a training set and a blank dictionary, initializing the blank dictionary and the cycle number, and setting an upper limit of the cycle number and a reconstruction error threshold;

(22) sparse coding: obtaining a sparse coefficient of a training sample by using an orthogonal matching tracking algorithm;

(23) and (3) dictionary updating: updating each atom in the dictionary and the corresponding sparse coefficient: firstly, finding out corresponding index set numbers of all training samples using atoms in a dictionary; calculating by utilizing atoms and sparse representation coefficients to obtain an error matrix; finding out the index number corresponding to the error matrix; finally, singular value decomposition is carried out to update dictionary atoms and sparse coefficients;

(24) and (3) loop iteration: after one-time sparse coding and dictionary updating are completed, the cycle times are increased automatically, whether an iteration condition is met or not is calculated, the iteration is stopped if the iteration condition is met, and otherwise, the execution is skipped (22);

(25) and (3) dictionary output: and obtaining and outputting a coefficient dictionary.

The sparse dictionary is a space formed by a group of overcomplete bases, and common fixed dictionaries include a discrete cosine base dictionary, a wavelet dictionary, a Gabor dictionary, a Gaussian dictionary and the like, and sparse expressions are carried out on signals of specific categories, and expressed sparse coefficients cannot reach the optimal sparse degree. By dictionary learning of a specific class signal set, a sparse dictionary with minimized average reconstruction error can be obtained. The aim of sparse dictionary training is to find a dictionary D that makes x as sparse as possible^*I.e. solving the equation:

inputting: training set: fuse target echo sample Y ═ Y₁,y₂,...,y_N},y_i∈RⁿOr foil strip interference sample Y ═ Y₁′,y₂′,...,y_N′},y_i′∈Rⁿ(ii) a The upper limit Num of the cycle times; reconstructing the error threshold Err;

and (3) outputting: a fuse target sparse dictionary D/foil strip interference sparse dictionary D';

the training samples of the same research object under different scenes are used for learning the sparse dictionary to obtain different sparse dictionaries, and an ultra-complete dictionary library of the specific research object is constructed. The number of times and scenes of offline learning are enough to meet different combat environments, and the sparse dictionary has stronger reconstruction capability on related signals.

(30) Sparse representation: based on an orthogonal matching tracking algorithm, performing sparse representation on the fuze echo signal after the normalization processing by using a target sparse dictionary, and outputting a sparse coefficient; or sparse representation is carried out on the fuze echo signals after normalization processing by using a foil sparse dictionary, and a sparse coefficient is output;

as shown in fig. 3, the (30) sparse representation step of performing sparse representation with the target sparse dictionary includes:

(31) pretreatment: inputting a target sparse dictionary, a fuse echo signal to be detected and sparsity, and initializing cycle times, sparse coefficients, an index set and residual initial values;

(32) updating an index set: calculating the maximum component index by using the sparse dictionary and the residual error, and completing updating;

(33) residual error updating: calculating an index column in the sparse dictionary and a signal after sparse coefficient reconstruction, and performing difference on a fuse echo signal to be detected to obtain a residual error;

(34) and (3) loop iteration: performing self-increment operation of the cycle times, calculating whether the sparsity condition is met, if so, stopping iteration, otherwise, skipping execution (32);

(35) and (3) coefficient output: and obtaining and outputting a sparse coefficient.

The process of performing sparse representation on the normalized fuze echo signal by using the foil sparse dictionary is similar to the process of performing sparse representation by using the target sparse dictionary.

(40) And (3) threshold judgment: vector-based l₂Calculating to obtain a sparse reconstruction error by using a fuse echo signal, a sparse coefficient and a sparse dictionary according to a norm theory, and judging the type of the fuse echo signal according to a preset threshold value;

(50) and (4) outputting a result: and if the echo signal is the fuze target echo, directly outputting the reconstruction signal, otherwise, subtracting the echo signal from the reconstruction signal to obtain the fuze target echo.