Research Of Ultra-Wideband Communication Technology Based On Compressed Sensing

In any wireless communication system, one of the most desirable features userswant is high speed data transfer. However, most of the current wireless systems arelimited mainly by their respective bandwidths. Ultra-Wideband (UWB) has emergedas a promising technology for high-speed indoor short-range wireless communicationsystems. The major reason for UWB technology to receive much attention is itspromising ability to provide low-power consumption, high bit rate, multipathresolution, and coexist with the narrow-band system by trading bandwidth for areduced transmits power. Because of the characteristics advantages, UWB trulybecomes an attractive technique for wireless applications in the future.UWB receivers, however, also has many technical problems. As we all know, todigitize a UWB signal, a very high sampling rate is required according to Shannon-Nyquist sampling theorem, but it is difficult to implement with a single Analog-to-Digital Converter (ADC) chip. New approaches for UWB receivers are needed toattain the required sampling rates. It is a key issue to find new methods of dataacquisition and processing.New theory emerged recently Compressed Sensing (CS) is regarded as the bestsolution to reducing sampling rate in receiver of UWB communication systems. Usingcompressed sensing, a signal can be sampled at sub-Nyquist rate leading to a reducedsampling rate and, hence, to a reduced the hardware complexity of the system.In this thesis, our major contribution is to exploit the new CS theory in thecontext of UWB communication technology. The aim of this work is to solve thebottleneck of the sampling at the receiver that is not feasible with state-of-the-art ADCtechnology. Firstly, two approach based on CS for UWB channel estimation areproposed. It is considered in both time domain and frequency domain. Extensivesimulations demonstrate that both the proposed models can perform accurate channelestimation based on the algorithms with the sampling rate can reduce to less than10%of the Nyquist rate.In addition, during the implementation of channel estimation using the channel models proposed by the IEEE802.15.3a, it can see that multipath arrivals in an UWBchannel mainly depend on the channel environments that generate different sparselevels (low-sparse or high-sparse) of the UWB channels. According to this basis, wehave analyzed and chosen the best recovery algorithms which are suitable to thesparse level for each type of channel model. Simulation results demonstrate that thesignal can be recovered exactly with the Basis Pursuit (BP) algorithm. In addition, byusing BP algorithm, the sampling rate at the receiver can be reduced to a Sub-Nyquistrate. In particular, BP is less affected by the different channel environments. However,it requires a complexity that makes the procedure extremely complex andcomputationally slow compared with Orthogonal Matching Pursuit (OMP). Whereas,the OMP algorithm greatly affected by different channel models. In the high sparse ofchannel models, the OMP method performs very well, which gives both the lowsampling rate and the fast processing time. However, in the low sparse of channelmodels, OMP method requires a higher sampling rate than the BP. Especially, in thecase of the multipath arrivals extremely dense, resulting in very low sparse of channelmodels, both methods BP and OMP become extremely slower, sometimesunacceptable. Addressing these cases, a new greedy algorithm named Block sparsechannel-OMP (Bsc-OMP) is proposed. The basic idea is based on an extension ofOMP. However, instead of simply selecting the largest component of one column at atime, Bsc-OMP selects the block that is the best match to the current residua. By thisway, the number of iterations of the Bsc-OMP is less than with OMP and, hence,significantly reducing the runtime.The short impulse duration, rich multipath environment and long channel delayspreads are the main cause of Inter Multipulse Interference (IMI) in UWBcommunication systems. In order to analyze this IMI, a new receiver structure forUWB communication systems based on CS is also proposed in this thesis. Thisproposals focus on solving two following problems. Firstly, application of CS theoryto reduce the sampling rate of ADC at the receiver. Secondly, analyses of the impactof IMI on high-speed data transmission of UWB communication systems. Theanalysis results show that the proposed model can successfully recover signal at thereceiver with the sampling rate to less than12.5%of Nyquist rate. In addition, whenanalyzing the effects of IMI, it can see that IMI is heavily dependent on the different types of channel model. As for Line-Of-Sight (LOS) channel models, the impact ofIMI is negligible, whereas for Nonline-Of-Sight (NLOS) channel models are greatlyinfluenced by the IMI. Moreover, the results also present that the Bit Error Rate (BER)performances are significantly improved by using the receiver with the structure ofthe Decision Feedback Equalization (DFE).Finally, by exploiting sparse characteristics of the image after wavelet transform,combined with pre-coding filter and UWB channel itself as a part of the measurementmatrix, an actual experiment to transmit an image via UWB wireless channel based onCS has also been performed. The simulation results present that in the case ofconsidering the effect of noise, BER performance for the proposed model issignificant improvement compared to the traditional model.