The netted radar has widely distributed radar nodes, which can deploy radar nodes by tasks and demands. It is a current research hotspot. This paper aims to study different beam pointings and predict the one-dimensional array node combination through neural network to reduce the sidelobe of the beam pointing. The dataset generated by genetic algorithm (GA) is fed to convolutional neural networks (CNN) for training, and the trained CNN can quickly predict. In this paper, a node selection algorithm based on GA-CNN for reducing array sidelobe is proposed, which combines the advantages of GA in dealing with combinatorial explosion problems and deep learning which has good generalization ability and prediction speed, and searches and predicts in the full set. It can be seen from the simulation results of the linear array that the CNN has learned the partial correspondence between the beam pointing and the node selection, and the operation speed is greatly improved. It makes the radar have further research value in the efficient adaptable environment.

The networking system consisting of multiple-input multiple-output (MIMO) radars, which inherits the merits of both MIMO radar and networking system, has important research value. In the MIMO radar networking system, due to the flexible and controllable system resource, the radar nodes in the network can cooperate with each other to complete certain tasks according to the operational requirements. Aiming at the background of multi-target tracking application, this paper discusses how to adaptively determine the working time, the radar nodes of opportunistic cooperative target tracking, the transmitting energy of the activated radars and waveform types and parameters of each activated radar. Firstly, the cooperative target tracking optimization model is established, in which the objective function considers the resource consumption and the multi-target tracking accuracy comprehensively. Then, the opportunistic cooperative target tracking algorithm is obtained by solving the established model via the particle swarm optimization. Simulation results show that compared with the non-cooperative and other fixed-parameter tracking methods, the comprehensive cost of the proposed tracking algorithm is the lowest, which verifies the effectiveness of the proposed algorithm.

Aiming at the problems that the characteristics of non-cooperative air targets are difficult to measure and the verification of imaging algorithm is difficult, the distributed radar ISAR imaging simulation system is constructed with the multi-angle ISAR imaging fusion algorithm. CST simulation software is used to obtain the complex scattering coefficients of typical air targets. Then the simulation constraints are studied. Finally, the distributed ISAR imaging simulation is carried out for a bomber with good stealth performance. The simulation results show that the distributed radar ISAR imaging simulation system can obtain the high-resolution ISAR target images at any observation angle.

A robust track initiation algorithm based on multi-algorithm fusion learning is proposed to resolve the correct and effective track initiation issue due to the effect of clutter and jamming in strong ECM and complex radar mission environment. This method regards the track initiation issue as classification issue, and uses the classical machine learning classification algorithm—random forest and GBDT as basis for fusion and classification. Chair-Varshney optimal decision fusion is applied to these two methods to achieve the effective and correct initiation of the target track. The simulation is used to compare the proposed method in this paper with random forest, GBDT, and Heuristic rule. The results show that the robust track initiation algorithm based on multi-algorithm fusion learning has better overall performance, much better than that based on GBDT and Heuristic rule.

This paper uses the target orientation and positioning technology of floating weak electromagnetic radiation sources on the sea surface, which only passively receives the electromagnetic signal transmitted by the VHF band buoy through the airborne platform，and calculates the accurate position of the radiation source by using the position information of the airborne platform，it is realized buoy search and positioning. Through the research on the target orientation and positioning technology of floating weak electromagnetic radiation sources on the sea surface, this kind of targets can be positioned quickly and accurately. It is one of the effective means to curb the detection equipment such as sonar buoy and wave boat.

For airborne bistatic cooperative detection radar, owing to the deployment and movement of airborne radar platform, the characteristics of clutter for bistatic radar are more complicated than monostatic radar. The space-time distribution characteristics of airborne bistatic radar clutter is analyzed. Based on the traditional airborne radar clutter model, a more accurate airborne bistatic radar clutter model is given, aiming at the geometric relationship of airborne bistatic radar and considering the relationship of the receiving beam and transmitting beam. The space-time distribution characteristics and the space-time two-dimensional clutter spectrum of airborne bistatic radar clutter in different airborne deployment and attitude situations are simulated, which provides technical support for clutter suppression of the airborne bistatic cooperative detection radar.

In order to analyse the relationship between the space-time distribution characteristics of clutter under the airborne bistatic radar system and the bistatic configuration, the geometric model of bistatic airborne radar is established, and the analytical expression of clutter isorange contours is derived. Then, a modeling method based on the variables of bistatic range sum and azimuth angle of clutter point relative to the center of the clutter isorange contours is presented to accurately calculate the Doppler frequency of clutter in any bistatic configuration. Simulation results show that the correlation of clutter range becomes stronger with the decrease of the bistatic range sum. However, with the decrease of bistatic baseline, the range dependence of clutter is relatively reduced. Therefore, the range dependence of clutter can be reduced by rationally configuring the bistatic parameters, which provides a theoretical basis for the suppression of bistatic clutter.

In order to solve the problem of long development cycle of traditional FPGA software in the realization of wideband channelized receiver, a fast implementation and verification method of polyphase filter channelization algorithm based on SysGen is proposed. The design idea of modularization and parameterization is used to complete 50 % overlapping polyphase filter channelization processing function, which has large bandwidth real-time processing ability and multi-signal receiving ability. Compared with the traditional programming development mode, this method can significantly shorten the development cycle and improve the flexibility of module development,realize wideband multi-signal real-time processing of FPGA channelization module, which can meet the requirements of contemporary wideband digital receivers.

Facing the increasing all-round and all-day air-atomic threat, in order to make up for the limitations of traditional radar networks deployment in dealing with jamming and insufficient three-dimensional coverage effect, and to improve the ability of detection and survival, a genetic algorithm based method of optimal deployment for three-dimensional radar networks under jamming situation is proposed in this paper. A radar range model under jamming situation is established and the radar deployment is optimized from three aspects: coverage emphasis, coverage capability and coverage overlap. Using this model, combined with genetic algorithm, the reasonable allocation of radar resources in protecting important areas can be realized, which provides an idea for building an omni-directional, three-dimensional and multi-layer protection network.

For the towed jamming indistinguishable problem of active seeker or airborne guidance radar, the towed decoy jamming mechanism is analyzed in detail. From the perspective of signal distribution characteristics and Doppler spectrum characteristics of target and decoy, using complex monopulse ratio and variable scale discrete Chirp-Fourier transform, a simple and effective towed decoy detection and identification technique is proposed to separate the towed decoy from the target. The validity of the algorithm is verified by computer simulation. The simulation results show that the time-frequency analysis and processing method based on variable scale discrete Chirp-Fourier transform has strong discrimination ability against towed jamming under some special conditions. The research results provide a way for tracking and guidance radar to counter towed jamming.

In view of the complexity of radar waveform and the decline of radar emitter signal recognition accuracy based on conventional pulse characteristics, a network structure of double CNN is proposed to realize the classification and recognition of 9 common radar signals. When using a single CNN structure, four modulation types can be identified accurately, but the recognition accuracy of phase coding and its composite modulation signals is intolerable . Because the time-frequency characteristics of BPSK and QPSK are similar. This paper adopts the processing method of double CNN structure, which has strong adaptability. Radar signals can still be classified and recognized when modulation parameters are not fixed. The simulation results show that the recognition accuracy of 9 modulation signals is higher than 95% when the signal-to-noise ratio (SNR) is 0 dB. Finally, the reliability of this method is verified by analyzing the relationship between recognition accuracy and SNR.

Marching vehicle-borne radar enjoys superiority in flexibility and agility, which conforms to the trend of radar development in the future. However, moving vehicle may result in frequency deviation of ground clutter from zero, and also lead to Doppler broadening of ground clutter spectrum. Consequently, the performance of conventional clutter suppression method dramatically declines. To cope with such difficulty, this paper presents analysis of clutter characte-ristics in marching vehicle-borne radar, and proposes a suppression technique based on space-time adaptive processing (STAP). The effectiveness of the proposed method is verified by simulation.

Aiming at global navigation satellite system (GNSS) spoofing signal recognition and intelligent anti-spoofing, this paper proposes a method of GNSS spoofing signal identification, spoofing signals orienting and adaptive nulling, based on estimation of signal parameters via rotation invariant technique (ESPRIT) algorithm and generated-signal sample matrix inversion (GS-SMI), combining the spatial spectrum estimation technology and adaptive array processing technology in array signal processing. The results of the simulation indicate that the proposed method is effective to restrain the influence of the GNSS spoofing attack. It provides theoretical support for further application of array antenna in GNSS spoofing environment.

For the issues of the low degree of approximation and incomplete elements of the test electromagnetic environment construction in the anti-jamming infield test of radar guided weapons, based on the analysis of the complex electromagnetic environment and the characterization surface test electromagnetic environment faced by radar guided weapons, the equivalent construction method of the dynamic test electromagnetic environment based on the characte-rization surface was presented. It focused on the dynamic test electromagnetic environment equivalent construction from three dimensions: combat scene mapping, combat technology equivalence and confrontation tactical control. It used the infield hardware-in-the-loop simulation platform to analyze the specific implementation method and the confrontation intensity index of the test electromagnetic environment was analyzed from the level of confrontation intensity. The method can provide the method support and theoretical basis for the test electromagnetic environment construction of anti-jamming test of radar guided weapons.

Aiming at the interference penetration of the jammer in the case of one-jammer to multi-radar, a reinforcement learning-based interference resource allocation method in the case of one-jammer to multi-radar interference is proposed. The interference radiation energy ratio and penetration distance ratio are introduced as evaluation indicators, and the dynamically adjusted reward values are used for DQN (deep Q network) and Dueling-DQN algorithms to enhance the convergence ability of the algorithm. By building a one-jammer to multi-radar interference penetration scenario, DQN and Dueling-DQN algorithms were verified, the experimental results verify the feasibility and difference of the two algorithms, and realize the resource allocation ability for interference resources in interference power, duration, interference pattern and interference radar selection, and meet the real-time and dynamic interference resource allocation requirement in the case of one-jammer to multi-radar.

In order to solve the problem of low recognition accuracy due to improper selection of signal characteristics in the process of jamming recognition by radar in modern electronic warfare, a jamming recognition method based on RBF neural network combined time-frequency domain analysis is proposed. This method proposes to combine the time-frequency domain characteristics of the jamming signal and take the advantage of the fast convergence speed and strong nonlinear fitting ability of the RBF neural network to improve the recognition probability of radar against active suppression jamming. The simulation results show that the RBF neural network based on time-frequency domain analysis can guarantee a high jamming recognition probability.

Aimingat the problem that the traditional method(chi-square) has poor ability to identify false targets when the angle difference between the radars relative to the target is small, an adaptive anti-rangedeception jamming method combining chi-square test and support vector machine is proposed. Firstly, determine whether further screening is required according to the number of combinations that meet the chi-square test; If further screeningis required, use the Mahalanobis distance(chi-square test variable), Euclidean distance, and Randolph distance between tracks as the multi-feature input of SVM. Then, train the SVM model. Finally, use the trained SVM to identify true and false tracks. The simulation results show that the true target track recognition rate of this method is 95.5%, which is better than the chi-square test and the improved chi-square test method significantly. In addition, the training time of this method is short; the recognition time is equivalent to the traditional method; the real-time performance is good, and the engineering application is strong.