| Traditional sampling method must obey Nyquist sampling theory, which states that the sampling rate should be twice higher than the maximum frequency of the signal. With the development of communication technology, the traditional sampling method brings increasing pressure on signal acquisition and processing devices. Recently, the theory of compressive sensing(CS) is proposed, it can be sampled the sparse signals at an extremely low frequency and recover the signal accurately. So the system storage and transmission pressure are greatly reduced.For analog signals, the sampling method based on CS is named Analog to Information Conversion(AIC). AIC can sample the sparse signals at sub-Nyquist frequency by configuring the suitable analog front-end structures. Within the AIC structures proposed recently, the Modulated Wideband Converter(MWC) suitable for multiband signals has broad application prospects. In this paper, focusing on MWC system, the main study are as follows:1. The research of MWC principles. For the MWC is based on CS theory, firstly the basic theory of CS is introduced. Secondly, the system constitution and working principles are described. In ideal situation, the analysis of frequency domain is introduced in detail, then obtaining the conditions that system parameters need to meet for ensuring the accurate recovery. Thirdly, the signal recovering algorithm especially the Simultaneous Orthogonal Marching Pursuit(SOMP) recovering algorithm is introduced. Finally, through the simulation experience, the design method and system performance are tested2. The study of influence factors of MWC system and design of MWC platform. The influences of system parameters, such as the number of channels, extending factor and analog low-pass filter, are studied. Then the optimal parameters of MWC system are determined. According to the structure design and parameters calculation, the MWC system platform is built which is composed of software part and hardware part. The mixing-filtering module is composed of the proper mixer, amplifier, high-order passive low-pass filter and so on; the four channel m sequences is produced by FPGA and then adjusting the output signals by subtracting circuit; The signal sampling, transmission, saving and dealing are completed by PXIe bus equipment. For software part, the relationship and program flows are introduced. The software part is based on Lab VIEW which can achieve waveforms displaying, data storage, parameters setting and algorithm adjusting and so on, the human-computer interaction is good.3. The research of actual MWC system sensing matrix calibration method. In the actual MWC system, the actual elements are different from the ideal simulation model, such as the non-linear property of mixer, the non-ideal property of low-pass filter and so on. These non-ideal factors affect the reconstruction effect of the system. Firstly, the possible non-ideal factors of the MWC system are analyzed. Then, a sensing matrix calibration method of actual MWC system based on sinusoid signals is proposed. By the proposed algorithm, the sensing matrix of the actual system can be exactly obtained which can be used to recover the location and the width of the spectrums.4. The experimental verification in actual MWC platform. Hardware system experiments based on the platform and calibrating method are made. The results show that the proposed MWC system and corresponding recovering algorithm can sample the multiband signals at a 12.8% compression ratio, and accurately recover the location and the width of the spectrums. |
PDF Full Text Request