In this thesis work, we first analyze the performance of the data fusion CFAR (Constant False Alarm Rate) detector and propose its modified type. And then we develop a programmable radar signal processor for realtime radar signal processing. In general, the radar with multiple receivers yields better detection performance than that with a single receiver. The data fusion CFAR detector integrates information from each receiver and gives decision on the presence of a target while keeping a desired false alarm rate. Under the homogeneous situation and the Rayleigh target model we can obtain superior performance with data fusion CFAR compared to the single receiver. But it has several problems. One of them is that the optimal solution varies if a target has different signal to noise ratio. With the proposed modified data fusion CFAR detector, we can overcome the problems with slight loss of detection performance. In the development of the programmable radar signal processor we use two ADSP-21020 digital signal processor chips and one FFT processor chip. It is hard to process the radar signal real-time because of its high data rate. We propose a modular structure to develop a real-time radar signal processor. With several radar signal processor modules one can achieve real-time radar signal processing.
