From the waveform point of view, radars are divided into two main groups: continuous wave (CW) and pulse radars. In CW radars, the transmitter is always on; in pulse radars however, the transmitter is on intermittently. In this thesis, we focus on CW radars. CW radars are often used for short range applications. Preliminary CW radars are capable of measuring the velocity of targets, but not the range of the target. To overcome this problem, frequency modulated CW (FMCW) radar was proposed. The transmitter of an FMCW radar sends a chirp signal; if a target is present, a shifted version of the transmitted signal enters the receiver where it is first amplified and then further guided to the mixer block. The mixer block multiplies the transmitted signal by the received signal. Then, ideally speaking, the output of the mixer would be multiple sinusoids. The number of the sinusoids and their relevant frequencies are proportional to the number of targets, and the range of targets, respectively. In practice however, due to disturbances such as noise and clutter, the output of the mixer is not pure multiple sinusoids but rather multiple sinusoids buried in noise. For this reason, the output of the mixer is sent to a detector block whose task is to determine the number of targets statistically. Once the number of sinusoids are detected, the estimator block estimates their corresponding frequencies in order to estimate the range and velocity of the targets. Regarding the estimation part of the system, we compare a new estimation method based on the combination of fast Fourier transform (FFT) and the chirp-z transform (CZT) with the conventional MUSIC method. The main focus of the thesis is on the detector block of the FMCW radar. The most well-known detector in FMCW radar is an intuitively proposed adhoc detector. Focusing on the detector block, we model the problem as a multiple composite hypothesis test. Then we propose a multi-step decision rule, which is based on the generalized likelihood ratio test (GLRT). We also apply the proposed detection strategy for the case of moving targets, which is tackled by a simple signal processing technique. The performance of the proposed detector is then compared with the conventional method through simulations. The simulation results reflect the effectiveness of the proposed method in both scenarios. Keywords Radar, FMCW, Detection Theory, Estimation Theory
