Detection and classification of electronic circuits with harmonic radar

Nowadays, the vast majority of electronic devices contain non-linear circuit elements such as diodes, transistors, semiconductors, etc. These nonlinear circuits cause re-radiation at the harmonics of the transmitted frequency. Thus, harmonic radar is used in the detection and classification of these circuits by Radio Frequency (RF) broadcasting. While the majority of the studies in the literature are based on detection, ranging, and tracking of electronic devices, a limited number of studies have focused on distinguishing and classification of electronic devices with harmonic radar. In this thesis, re-radiated signals from nonlinear electronic circuits were analyzed and unique methods were developed for the detection and classification of these circuits. In this thesis, firstly, a novel method for classifying electronic circuits with a Frequency-Sweep Harmonic Radar (FSHR) approach is presented. Unlike studies in the literature, a frequency-swept signal with a constant power is transmitted to the electronic circuits. Different types of electronic circuits exhibit different harmonic responses, while the same type of circuits show similar harmonic behavior via FSHR approach. Statistical and Fourier features obtained from harmonic responses are utilized to classify electronic circuits. Classification performance is expressed in both harmonic and feature spaces. The harmonic space is defined as a three-dimensional space consisting of the first three harmonics. On the other hand, a three-dimensional feature space consists of statistical and Fourier parameters. Furthermore, the classification performance of the FSHR approach is evaluated with Monte Carlo simulations using harmonic responses in the presence of noise. In the second study, as a continuation of the first study, new feature sets are used to classify electronic circuits. The nonlinear responses of the circuits are recorded by transmitting constant power frequency-swept signals to the electronic circuits. Unlike the previous work, statistical and Fourier feature sets created from both harmonic and feature parameters. Rather than analyzing the feature parameters and harmonic information individually, using all parameters as a feature set provides a good basis for performance comparison. In another study conducted within the scope of the thesis studies, a new approach is developed for the classification of electronic circuits with harmonic radar. Unlike previous studies, harmonic responses of electronic circuits are associated with parameter vectors with the proposed linear model. Power-swept time-varying single-tone and two-tone signals are sent to electronic circuits and distinctive parameter vectors are obtained for each circuit. In the presence of Complex White Gaussian Noise (CWGN), estimated parameter vectors are achieved using the Maximum Likelihood Estimator (MLE) and various electronic circuits are classified using statistical features of the normalized estimated parameter vectors. The simulation results show that the parameter vectors successfully classify the electronic circuits with the presented linear model method.