Digital Pulse Compression System Real-time Processing Software Design

Pulse compression technique is widely used in modern radar systems. On the premises to ensure the resolution in speed domain, large timeband product signals are used in pulse compression technique to improve the resolution in range domain. This dissertation mainly studies the radar multi-waveform digital pulse compression processing system in frequency domain, analyzes the pulse compression performances of linear frequency modulation signals (LFM), nolinear frequency modulation signals (NLFM) and Taylor quadriphase code signals, designs the side-lobes suppression filters, and develops a multi-waveform digital pulse compression system based on four pieces of ADSP21160. The system can process LFM, NLFM ,Taylor quadriphase code signals with time width under 53 μs and band width under 5MHz. The experimental results demonstrate that the specification characteristics meet the requirements of practical radar systems. The main content and creative work in this dissertation include: (1) The characteristics of the LFM signals with band width equaling 5MHz, time width in 9μs~42μs, NLFM signals with band width equaling 5MHz, time width in 9μs~21μs, Taylor quadriphase code signals with band width equaling 3.3MHz, time Width in 2.7μs~7.2μs are comprehensively investigated in frequency domain pulse compression. The corresponding side-lobes suppression filters are designed. (2) The structure of the digital pulse compression processing system is discussed; the pulse compression program of 4 pieces of ADSP21160 in the process board, the program of single piece of ADSP21160 and the time logic of CPLD in the interface board are developed and optimized, and the system debugging work is completed. (3) When Taylor quadriphase code echo signals occur at the overlap of the segments of cyclic convolution, a special high side-lobe is found in the output of pulse compression during the course of system debugging. The cause of the side-lobe is analyzed based on the theory of cyclic convolution and a corresponding suppression method is presented. After processed, the echo signals at any point have the same RMS which is larger than 30dB. (4) The performances of the system are analyzed based on the experimental data. The RMS of signals with time width over 10μs is larger than 34dB, the RMS of signals with time width under 10μs is less than 30dB. When the SNR of input signal over 70dB,...