IR-UWB Communication Performance Research Based On Compressed Sensing

IR-UWB technology is an excellent wireless communication technology, ithas many technical advantages. But in the traditional IR-UWB communicationsystem, high-speed sampling unit is difficult to implement for low-cost andlow-power. In recent years, compressed sensing is able to solve the technicalbottleneck faced by IR-UWB, which attracts wide attention.This thesis takes full advantage of the sparse characteristics of the IR-UWBsignal and multipath channel. The performance of IR-UWB communication basedon compressed sensing is studied deeply. At first, introduce the basic principles ofthe IR-UWB and compressed sensing, and then demonstrate the feasibility ofapplying compressed sensing to the IR-UWB system. IR-UWB communicationsystem based on compressed sensing is proposed. On the basis of thiscommunication structure, the reception performance of the IR-UWB system isstudied, in which communication method is data-stream or burst. Then combinewith new signaling scheme, new channel estimation and signal demodulationalgorithms are proposed based on the feedback architecture. Its performanceenhance3~5dB as compared with the ordinary channel estimation in the samerelative error. Meanwhile, the performance of signal demodulation algorithm basedon the feedback architecture gradually approach to the performance of digitalreceiver with the full sampling rate with the signal-to-noise ratio increasing. Itsolves the problem that the performance of ordinary signal de-modulationalgorithm based on compressed sensing differs largely from digital receiver withfull sampling rate, and enhances the performance of IR-UWB receiver based oncompressed sensing. When compressed sensing is used to the reception of burstIR-UWB signals, firstly attributed it to the blind de-convolution problem under theframework of the compressed sensing, then establish the model of solving thisproblem, finally the signaling scheme of burst IR-UWB communication is designedon the basis of the model. To solve this problem, the existing algorithms are studiedand find their computing complexity is very high. In order to reduce thecomputational complexity, and keep similar performance, the improved algorithmis proposed. The simulation results show that the improved algorithm is correct.