Researches On The Techniques For The Sub-Nyquist Theory And Applications Based On MWC

With the development of electronic technology,vrious transmitters have arised,and the signal environment is more complex with that the frequency distributes in a large range,and the bandwith of the signals which needs to be monitored are larger and larger.For the traditional Nyquist sampling theory,high sampling rate and big data size have become a main factor that restricts the non-cooperative wideband signal sampling and digital processing.The fact that sampling rate is too high and data is too large has presented a huge challenge to the existing analog digital converter(ADC),digital signal processor(DSP)and memory devices.Sub-Nyquist sampling technology based on Compressed Sensing(CS)is an important research topic in signal sampling and processing field in recent years.It uses the fact that wideband signals are sparse in frequency domain or other transform domain to realize wideband signal sampling and perfect recovery with sampling rate far bellow Nyquist rate.One,named modulated wideband converter(MWC),which has low request about hardware and the signal model matching degree,has presented huge advantages in sub-Nyquist sampling technology and has drawn much attention from scholars.Currently,the study of MWC sampling technique is still on the way which concentrates on the four contexts of the impacts of non-ideal factors and the resolutions,the improved MWC structures,sensing and reconstruction technology,applacations et al.However,there are still problems and shortcomings,the technique needs further study.This paper is aiming to investigate the following four problems in MWC technology: 1)power spectrum sensing and reconstruction with single channel MWC structure;2)wideband signals specific selective sampling;3)burst signal and other non time continuous signals sampling in wideband spectrum;4)wideband spectrum fast sensing.The main contents of the dissertation are as follows:1.Aiming at the problem that MWC structure is complex,the research on the power spectrum sensing and reconstruction based on single channel MWC structure is studied.Firstly,a single channel MWC sub-Nyquist sampling structure has been designed which has low complexity and can be realised by available devices;moreover it has reduced the total sampling rate of MWC significantly.Then,according to the characteristic of Wide-Sense Stationary(WSS)signals that the power spectrum does not vary with time,a relation between the power spectrum of wideband signal with the Fourier transform of samples from proposed scheme,and an underdetermined sparse function is constructed.At last,a method of power spectrum sensing and reconstruction is presented based on the unique sparse resolution of measurement function.Compared with MWC,the proposed structure has lower hardware complexity.Simulation results verified the effective of the power spectrum sensing and reconstruction scheme based on the proposed structure.2.Aiming at the problem that there are interfences of disinterested bands,two specific bands selective sampling method of wideband signals are studied.On one hand,according to the processing after sampling,a selected sampling method based on orthogonal projection is proposed.It uses orthogonal projection to construct a linear selective matrix to operate on the sub-Nyquist samples from MWC in order to filter out disinterested band information while guarantee the signal in specific bands can be reconstructed perfectly from the left information.On the other hand,according to the front-end design,a selective sampling method based on periodic waveform redesigning is proposed.It redesigns periodic waveform according to the band information based on sign periodic waveform,to enchance or inhibit band selectively.Therefore,the selective sub-Nyquist sampling can be achieved with sparse wideband signal.The simulation results show the effectiveness of the two selective sampling methods.3.Aiming at the lower sensing and reconstruction performance with burst signal and other non-continuous time signals at low signal noise ratio(SNR),a method of sigal number,support and base band waveform union estimation based on sparsity adaptive is proposed.First,this method fully utilize the sparsity in frequency domain to realize the power and base band waveform union estimation of each band,and according to the power to estimate signal number and support.Then,according to the power variance compared with previous measure period and energy static characteristic of corresponding base band waveform to detect burst signal.Compared with original MWC reconstruction method,it has overcome the prior knowledge of band number,and can adjust to the situation that the support changes with time.Simulation results show that the proposed method has improved the performance of waveform recovery of burst signal and other non-continous signals and can adjust to lower SNR.4.Aming at the problem of large computation of existing MWC sensing method,a wideband spectrum fast sensing method based on MUSIC is studied.First,the MWC sensing problem is translated into the unique resolution of a MMV problem,and MUSIC is introduced into the MMV problem.Then,aiming at the shortcoming of MUSIC when solving MMV,a modified factor is introduced to improve MUSIC to inhibit the noise in MUSIC spectrum while enhancing the value where the signal is.and the sensing performance is imporved.At last,it has been compared with OMP sensing method in computation,and is been proved that its computation is small and don't increase with signal number increasing.Simulation results verify the proposed method can complete spectrum sensing fast and effetively.This paper extends the research MWC sub-Nyquist sampling techniques,explores some issues,such as reducing the hardware complexity.Some effective methods are proposed to provide the new means and research ideas for MWC sub-Nyquist sampling theory and applications.