Research On Technologies Of Cognitive Radar Probe In Complex Background

The “Four Countering” has still been a huge challenge for modern radar.As the scene is more complicated than before,its characteristics of non-stationary,dynamic and time-varying also brings higher requirement for radar detection.Radar has to deal with environmental clutter,frequency multiplexing,and also the issues of ?Four Countering?.Traditional radar with its limited mechanism lacks the information feedback from the receiver to transmitter,which makes it difficult to adjust waveforms and interact with outside,and also disable to provide enough probing performance.Cognitive radar(CR)as the novel system equips the feedback from receiver to transmitter to achieve waveform diversity.To deal with ?Four Countering?,its waveforms must consider the sidelobe masking as well as spectrum compatibility and cross-interference.Meanwhile,as hostile moving targets always impact on radar safety,how to combine target detection with anti-sidelobe masking and anti-intercepting together has been an urgent issue for CR.In addition,maneuvering targets often exhibit dynamic and random characteristics,which incurs maneuvering model mismatching.How to ensemble multi-model information fusion with waveform design has also been an obstacle.Aiming at issues above,this project focuses on the theory innovation in sophisticated battlefield,and further explores technology applications and strategies for ?Four Countering?,by using ideas of knowledge-aiding,waveform diversity and multi-model information fusion.This project would enrich the CR probing theory,and provide theoretical support for developping adaptive capability and detection technology of modern radar.First of all,aiming at sidelobes masking,cross-interference and active-interference when detecting static targets,we utilize the prior-knowledge aiding idea to tackle these problems.Based on alternating projection,the correlation optimization problem could be transformed into the frequency-norm approximation problem,and a multivariable objective function is further deduced in detail.The relaxation factor as well as acceleration factor have been introduced to enhance local optimization for nonconvex problem.By all these,we put forward methods of anti-sidelobe masking,anti-cross interference,and joint optimization of anti-sidelobe and spectral congestion based on spectral approximation relaxation alternating projection.Moreover,we further propose the dual-projection-based waveform design algorithm for anti-cross interference,anti-sidelobe masking,and spectral congestion.Meanwhile,considering the timeliness and convergence of gradient algorithm,a collaborative optimizing algorithm has been also introduced to design cognitive waveforms.Simulations show that these proposed algorithms have better performance than conventional methods,such as ISAA,CAN and gradient ones,which with stagnation,poor robustness and long-time dilemma.Secondly,aiming at poor Doppler tolerance,high sidelobe and poor anti-interception ability when detecting enemy moving target,a LFM-noise composite waveform is constructed where the phase modulation factor is used to connect the phase section of LFM and noise one.Then we introduce the cognitive multi-level dynamic template and non-dynamic templates by using prior knowledge,and present a novel waveform design algorithm based on the cognitive multi-level phase-modified relaxation alternating projection,to design waveforms with high Doppler tolerance,low sidelobe and low interception property.In addition,under Majorization-Minimization framework,we introduce the novel waveform design method for optimizing specific range-Doppler interval in fuzzy function map by using a two-dimensional mapping template.The range-Doppler sidelobe suppression could be transformed into a mathematical problem to obtain the desirable composite waveform.Simulations show that compared with several existing algorithms,the proposed algorithms could achieve better performance of high Doppler tolerance,low interception and low sidelobes.Thirdly,aiming at maneuvering target tracking under the case of maneuvering model mismatching,we propose a novel adaptive waveform design method based on matrix-weighted interacting multi-model fusion.The tracking framework is formulated by using matrix-weighted multi-model fusion in lieu of the probability-weighted way.Considering the Markov effect of model transferring,we propose an adaptive transition matrix updating mechanism to tackle this slowness.As the fused covariance matrix of state estimation selected as the ellipse metric,we rotate the measurement error ellipse to make them orthogonal to each other,and obtain adaptive waveforms.Simulations show that compared with existing algorithms,the proposed algorithm has good stability and small tracking error,which could meet the requirements of maneuvering target tracking.Finally,aiming at observation loss or pollution in the receiver of CR,observation and their virtual ones are constructed from the predicting idea and also as the preprocessing.The feedback of CR is improved by using the integrated data fusion and multi-model state fusion.Two novel waveform design method based on virtual observations fusion and probabilty-weighted or matrix-weighted multi-model fusion are proposed.Virtual observations fusion could be used in linear or nonlinear pre-filters,so that the multi-model information fusion and predicting feedback of receiver are combined to adjust the transmitting waveform.Simulations show that the proposed algorithms have obvious performance,and could meet the requirements of CR tracking.