Reseach On Key Technique For Spectrum Analysis Of Broadband Signals And System Implementation

The technology of signal spectrum analysis based on digital processing features high measuring speed and real-time processing. For a long time, it has been the research hotspot of the scholars and engineers both at home and abroad. However, with rapid development of the modern science and technology, in some fields, the digital signals present the features of complexity and variety. These features are embodied in the widening signal bandwidth, the growing carrier frequency, the incearsing instantaneity and complexity, and the rapidly rising nonstationary, etc. It becomes difficult for the analog or digital measuring system based on traditional hardware technology and soft architecture to finish the rapid analysis and test of these types of signals.Analysis and measurement of the spectra of the broadband signals needs the research breakthrough of the high-speed and high-accuracy data aquisition and the technology of broadband digital intermediate freqeucny(IF) processing. The increasing bandwidth and carrier frequency of the digital signals request the high sampling speed and accuracy of the analog-to-digital Converter(ADC) in varieties of digital signal analyzing and mearsuring equipment, meanwhile, the high speed of the ADC output puts huge stress on the follow-up digital signal processing(DSP) chip and the whole signal measuring system. How to implement the high-speed and high-accuracy data acqusiztion, and then the spectrum analysis of the input broadband signals is the emphasis of this dissertation. Combining the related projects during the period of pursuing the doctor’s degree, and aiming at the key techniques of sampling and spectrum analysis of the broadband signals, this dissertation does the deeply researches on the following aspects:(1) Time-interleaved ADC(TIADC) parallel sampling system. The channel mismatch error model is established. Based on this model, the frequency-domain expression of the TIADC output with the linear channel mismatch errors(offset, gain and timing mismatch errors) is derived, and the impacts of these mismatch errors on the TIADC output are analyzed. Two simple and easily-implementable foreground channel mismatch error estimation algorithms are proposed: one is based on the Discrete Fourier Transform, and the other is based on fractional delay and sine curve fit. The correctness and effictiveness of the two algorithms are verified by the simulation results, and the estimated errors are corrected and compensated by the fractional delay filter based on the FARROW structure.(2) ADC parallel sampling system based on hybrid filter bank(HFB), or frequency-interleaved ADC(FIADC). The perfect signal reconstruction condition in this system is analyzed and derived, as well as the frequency-domain expression of the system output with the existence of the channel mismatches. The impacts of channel mismatches on system sampling out are analyzed.(3) Technique of digital down converter(DDC). Based on the analysis of the speed restrict of the traditional serial DDC structre’s processing the high-speed data output by ADC and the signal polyphase expression, several parallel DDC structures with high efficiency are derived, including the postpositiona mixer structure, least common multiple structure, double frequency transfer structure and so on. These structures can relief the processing speed stress of the digital signal processing chips significantly.(4) Technique of numerically controlled oscillator(NCO) implementation. For the several parallel DDC structures, the mathematical relations between the related NCO parameters of the parallel DDC structre and serial one are analyzed and derived detailedly. Based on the information of the sampling rate and bandwidth of the input broadband signals, the selection principles of related parameters to implement the multichannel NCOs in the parallel DDC structure are discussed.(5) Technique of broadband signal spectrum analysis based on channelization. Combining the theories of polyphase expression and DDC, the technique of broadband signal spectrum analysis based on channelization is proposed to divide the frequency band of the input broadband signal into several subbands. Apply the DDC and fast fourier transform(FFT) to these subband complex signals to get the corresponding spectra. Two channelization structures are derived to realize the division, i.e. polyphase DFT structure and weighed overlap-add(WOLA) structure. A kind of passband splicing method is used to managing the frequency-domain data of subbands.(6) Frequency measurement. The technique of frequency measurement with the sine input in the channelization structure is discussed. The frequency of the input sine wave is calculated by working out the baseband frequency of the channelized bansband complex signal and then referring to way of channel dividing. The corresponding calculating formula is presented.(7) Technique of restraining signal truncation error. In DSP, restricted by the calculating precision of some digital algorithms, the truncating of bitwidth of signal is necessary. However, the direct truncating operation on signal leads to the decline of the quality of the truncated signal. In this dissertation, truncating after adding dither to the original signal could restrict the truncation error and improve the quality of the truncated signal significantly.(8) Verification of the broadband signal spectrum analysis technique. The designing objectives, general structures and implementation of key algorithms are introduced, and the experimental testing results are presented.