Ultra-wideband Communication Systems, Channel Estimation Algorithm And Its Vlsi Implementation Of The Study

As a competitive technology for short-distance wireless communications, Ultra-wideband (UWB) communications has recently received unprecedented attention from both academia and industry. It is featured with high data rates, low power consumption, reduced complexity of transceivers and low probability of interception.Due to unique properties of UWB signals and fading channels, system design has to cope with many challenges. Among them, channel estimation serves as one of the most crucial issues, since system performance largely depends on the accuracy of channel estimation. In UWB communications, the power of transmitted pulse is limited, which results in insufficient energy capture at the receiver. Moreover, UWB channels are much more complicated than general wireless channels, and a large number of channel parameters should be estimated. Another problem is the high sampling rate because of the ultra-short signals. Hence, it is difficult to design and implement a fast algorithm with low computational complexity and high accuracy.In this dissertation, the principle of typical UWB transmitters is introduced firstly. The propagation characteristics of UWB channels is described. The principle of UWB receivers, especially Rake receivers, is analyzed. The importance of channel estimation for receiver performance is illustrated.Several traditional estimation algorithms are investigated, i.e., Maximum-likelihood (ML) based, Least squares (LS) based and Pilot waveform assisted modulation (PWAM) based. Through theoretical analyses and simulations, comparisons and conclusions are drawn with respect to the performance, complexity, bandwidth efficiency, etc.A modified LS (MLS) algorithm is proposed based on traditional LS algorithm. A narrow window energy detection process is introduced at the first stage of MLS. It efficiently suppresses the noise component, and thus improves the estimation accuracy. Furthermore, a maximum-likelihood method is adopted at the second stage, replacing the complex LS method. It remarkably reduces computational complexity, while exhibiting a nearly equivalent performance compared with traditional LS algorithm.VLSI implementation for MLS algorithm is carried out, including quantitative analysis, module design, and platform design. The result of FPGA verification is given at last.