Adaptive Waveform Design For Target Detection In Heavy Sea Clutter

Target detection in heavy sea clutter is one of the most challengeable detectionproblems. It is difficult to improve radar detection performance by traditionaldetection methods. It is effective for radar system to take advantage of theenvironment and target characteristics in complex ocean environment. The mainpurpose of this paper is to study the estimation problems of sea clutter and designof adaptive waveform through the estimated information.Firstly, the sea clutter signal process handling frame of the parameter estimationand signal design is proposed. The Compound-Gaussian sea clutter model isanalyzed and temporal correlation properties are studied using IPIX radar data.Through SIRP method, clutter with appropriate temporal correlation properties isgenerated. After that, the processing procedure of the received signal is introduced.Then, the texture values and speckle covariance are estimated by EM algorithmand the statistical analysis of the estimation performance is made. Sea cluttersuppression based on principal component analysis (PCA) is achieved successfullyso that radar will predict the target location more accurately. The results show theeffectiveness of the suppression algorithm.Finally, a waveform design method by setting notches to the autocorrelationfunction is proposed to mitigate the effect of range side-lobes to the detection cell.The gradient vector and Hessian matrix belong to the second derivative-basedoptimization algorithms are derived. The maximum likelihood estimation of targetpower is obtained and GLRT detector is derived. The detection performance isobtained after the PCA clutter suppression. In order to validate the advantages ofthe designed PM waveform, the detection performance between LFM and PMwaveforms is compared. In addition, the impacts of shape parameter and the notchnumbers are analyzed. Simulation results show that detection performance isimproved after PCA suppression. Compared with LFM waveform, the designed PMwaveform will bring better detection performance. The algorithm is more suitablefor spikier clutter. Notch numbers will have an effect on the algorithm.