FMCW LPI radar signal detection and parameter extraction methods based on Wigner Hough transform

With more Low Probability of Intercept (LPI) radars coming into use, the signals of interest are changing at a rapid pace. Electronic Support Receivers (ESR) currently in use are not well optimized for LPI radar detection. The techniques used by conventional ESRs either cannot detect LPI waveforms or they can detect them only from a range much less than LPI radars' instrumented range, and this gives LPI radar platforms a range advantage. ESRs used for LPI radar detection shall use; time-frequency (TF) signal processing techniques, correlation techniques and algorithms to overcome the advantages of LPI radars. LPI radars use different waveforms for different purposes of the radar. In this thesis we focused on Frequency Modulated Continuous Waveforms (FMCW). Wide-band linear FMCW is one of the most important and common waveform structures used to achieve LPI operations. It is important for ESRs to autonomously detect and analyze FMCW and other LPI modulations due to their time-critical mission. In this thesis, TF space and parameter space are used for FMCW signal detection and parameter extraction. TF space is derived by calculating the Wigner Ville Distribution (WVD) of signal samples and parameter space is derived by applying Hough / Radon (H/R) transform to TF space points that exceed TF threshold. This combination of WVD and H/R transform (WHT) spaces makes it possible to redefine the problem of detecting FMCW components in WVD as detecting the high intensity points in parameter space. The detection decision is given whenever the amplitude of any point in parameter space exceeds the detection threshold. WHT based methods in literature are known to be computationally expensive since they use the entire WVD space and calculate HT for all the possible chirp rates. LFM signals above or below certain chirp rates are useless to detect any kind of real targets. Calculating the HT for only a limited set of chirp rates and for WVD pixels exceeding the threshold will reduce the required processing power and complexity. In this thesis, new WHT methods that reduce the complexity yet keeping detection and parameter extraction performance above a certain level are proposed. The performance of proposed WHT based detection and parameter extraction algorithms is compared in terms of transform speed, parameter extraction and detection performance. Also, the effect of chirp rate on FMCW signal detection using proposed WHT based methods is analysed. Following the chirp rate effect analysis, digital chirp rate adaptation method is proposed to increase WHT based methods' detection performance. The analyses provided throughout the thesis show that the proposed methods are candidates that can improve ESRs automatic signal detection and analysis capabilities with their speed and performance.