Study On Key Technologies Of Radar Target Parameter Estimation And Tracking

The parameter estimation and tracking of the radar target always is hotspot issue for the radar signal processing and data processing.Especially,the parameter estimation based on pulse Doppler radar,DOA estimation based on the sparse array and the tracking of the weak and maneuvering target are the key and difficult technologies in the radar system.Combining the Chinese remainder theorem and random finite set,the two key technologies of radar target parameter estimation and tracking are being studied.The main contributions of this dissertation are shown as follows:(1)For the current methods facing with big computer complexity to solve velocity ambiguity,the closed-form robust Chinese remainder theorem is proposed.The proposed algorithm possesses less computer complexity due to closed analytic solution.It brings difficulty for the method based on Chinese remainder theorem to estimate Doppler frequencies of multiple targets,because the correspondence between the detected targets and the wrapped frequencies is unknown in advance.The proposed algorithm,which is based on amplitude aid using clustering methodology,solves the problem of Doppler frequencies estimation of multiple targets.The theoretical analysis and simulation results validate the effectiveness of the proposed algorithm for Doppler frequencies of multiple targets.(2)To improve the performance of the target detection,the radar always operates on a high pulse repetition frequency(PRF)to obtain wider free clutter areas in frequency domain.The use of a high PRF can,however,lead to range ambiguity problem in many case.The closed-form robust Chinese remainder theorem is proposed to solve the range ambiguity problem with less computer complexity.Meanwhile,the Chinese remainder theorem is difficult to apply into multiple targets due to unknowing correspondence between the target and measured range.The symmetry polynomial aided Chinese remainder theorem is proposed to achieve the range estimation of multiple targets.The simulation results verifies that the proposed algorithm can accurately and robustly achieve the range estimation of multiple targets.(3)Compared with the traditional uniform linear array,the sparse array can extend array aperture to improve the estimation precision of DOA and the resource resolution.The traditional DOA estimation algorithm is difficultly applied to the practical engineering due to the strict geometry of the array.The proposed algorithm can estimate the DOA by solving phase ambiguity using Chinese remainder theorem.To solve the problem of the unknowing correspondence between the targets and the wrapped phase,the Doppler aided Chinese remainder theorem is proposed to realize the DOA estimation of multiple targets.The simulation results demonstrate that the proposed algorithm significantly improves DOA estimation precision with less computational cost,especially in the DOA-closely-spaced targets case.(4)Based on the TBD measurement model,the cardinalized probability hypothesis density(CPHD)filter is proposed to realize the tracking of the weak targets in complex clutter background.For the non-linear and non-Gaussian multiple target tracking case,the sequential Monte Carlo(SMC)implementation of the proposed algorithm is given.Meanwhile,the likelihood functions of the different type of target echo model in K clutter and G0 clutter background are deduced.The simulation results demonstrate that the CPHDTBD filters based on the above likelihood functions can well realize the tracking of the weak target in the complex clutter background.Additionally,the efficient parameter estimation method based on the moment estimation is proposed.(5)Combing Delta-generalized labeled multi-Bernoulli(Delta-GLMB)density with labeled multi-Bernoulli density(LMB),an efficient method based on TBD measurement model is proposed to track multiple weak targets in sea clutter background.Under the Bayes framework,the proposed algorithm consists of two steps including prediction and update.In the prediction step,the proposed algorithm fulfills the prediction by the LMB form and utilizes the dynamic grouping technology to implement the match between the predicting densities and the measurement.In the update step,each group of the filter fulfills update by the Delta-GLMB form.More importantly,the conjugate prior property based GLMB does not hold on for the TBD measurement model in the update step.In order to solve this problem,a tractable principled approximation involving GLMB density is used to make whole Bayes filtering with labeled RFS to perform iteratively for each group.The simulation results demonstrate that the proposed algorithm can drastically reduce the computational cost albeit with slightly precision loss compared with standard Delta-GLMB filter.(6)An efficient implementation of the multiple-model generalized labeled multiBernoulli filter based on TBD measurement model(MM-GLMB-TBD)is proposed to track multiple maneuvering targets with low signal-to-noise(SNR).The joint implementation of the MM-GLMB-TBD filter is firstly obtained to eliminate inefficiencies in the truncation procedures of the original two-step implementation.The lattice-reduction-aided Gibbs sampling with flexible proposal is proposed to effectively truncate the filtering densities in the MM-GLMB-TBD filter with better exponential rate.The simulation results demonstrate that the proposed algorithm can drastically reduce the computational cost without the estimation precision loss compared with the original implementation based Gibbs sampling.