Research Of Methodology For Software Defined RF Receiver Based On Nonuniform Sampling

Software defined radio is a useful tool for implementing RF receiver since it is mainly cost effective. Two methods have been investigated in this thesis, which facilitate the implementation of software defined RF receiver from the Digital Alias-free Signal Processing (DASP) point of view.One is to use Weighted Periodic Nonuniform Sampling (WPNS) for processing the RF signals. The methodology of measuring the level of aliasing within the required range of frequencies is proposed. The optimal WPNS is found by searching a carefully selected subspace of feasible solutions and applying Lawson algorithm for weight calculation. We describe The improvement that has been made to the algorithm of searching the optimal sampling instants are described. It is shown that WPNS has better alias-suppression properties than traditional PNS.The other is to formulate the evaluating method of signal reconstruction which analyzes the signal reconstruction quality. Hence the quality of signal reconstruction of RF signals could be assessed. The proposed tool, reliability of signal reconstruction, provides a quantitative measure of the quality of signal reconstruction by using the layout of the sampling instants and the spectrum support function of the sampled signal. For each time instant, the reliability function reveals the scale of potential imperfections in signal reconstruction resulting from limited provision of signal samples, their distribution (such as gaps and clusters) as well as aliasing.In summary, this thesis focuses on research of periodic non-uniform sampling based digital alias-free signal processing to use non-uniform sampling for maximizing the suppression of the effects of aliasing. Hence, software-defined radio system designers no longer need anti-aliasing filter and are released from the limitation of IF filter. The accuracy of signal reconstruction and the optimal signal reconstruction of the RF signal are also investigated. The obtained results provide the necessary software, methods and means for the implementation of the non-uniform sampling-based software-defined radio receiver.