Research On Some Key Techniques In Impulse Radio UWB Receivers

In recent years, UWB(Ultra-wideband) is becoming a hot research subject as a new physical layer technology for short distance high data wireless connection with low consumption and low cost. At present, there are two widely accepted technologies for UWB: PULSE DS CDMA based UWB and MB-OFDM based UWB respectively. In this dissertation, some key reception techniques on PULSE DS CDMA systems are discussed, mainly including synchronizing, channel estimating and receivers, and multiuser detections etc. In the dense multipath channel of UWB, the signal power is spread into dozens or hundreds of paths and serious ISI (inter-symbol-interference) will be caused. How to achieve better performance in the low power and ISI environment is the main problem we will face to. Based on the previous works of other researchers, some new methods and algorithms involved synchronization and reception are developed, which are studied in details in the following chapter 3,4 and 5 respectively .1. Time synchronizing. In the slow time variant UWB channel, for the received signals have the periodical-correlation characteristics between symbols, a new parallel auto-correlation synchronization method is proposed in chapter 3. By this method, the receiver can acquire more power and make fast synchronizing possible. The simulations show that the timing acquisition convergence gets fast without complex channel estimation like RAKE receivers. But because the timing accuracy depends on the step of each parallel circuit, the complexity of the circuit will increase rapidly to meet the high resolution of timing. To solve this problem, a GLRT(generalized likelihood ratio test) algorithm is proposed. The symbol coming will bring the sudden change of the GLRT and this change will appear repetitively when the symbol coming periodically. By detecting the sudden and repetitive change of GLRT, the more accurate and fast symbol timing is acquired. This result is also proved by the simulations. In the final part of chapter 3, the main factors, including SNR, detecting threshold, active user number and so on, affect timing synchronization performance are also analyzed and discussed in details.2. Single user receiver and its performance. In the dense multipath UWB environment, the basic receiver structure is RAKE. In the chapter 4, we first summarize a ML channel estimation algorithm (which has been widely used and has good performance) and then derive the SRAKE receiver performance. Due to the influence of channel estimation, the acquired power will be impaired. To combat the degradation, an autocorrelate detection technique is investigated. The simulation shows that its performance is better than that of ML+RAKE, especially at the high SNR. Besides, for comparison, we study the other UWB receiver based on subspace algorithm. For the influence of the multipath, waveform distortion and ISI, the DS UWB signal' spread code will not be simply looked as its subspaces. We must estimate the signal's subspaces. To decrease the calculation complexity, the PASTd algorithm is given to update the eigenvalue and eigenvector of the received signals self-correlation matrix. According to the simulation results, this algorithm is more efficient in CM1 channel than in CM3, because the ISI will be more serious in CM3. ISI makes the estimation getting worse and causes the "floor effect" of BER.3. Multiuser detection. Multiuser receiver is very different from that of the single user, because the multiuser interference will degrade the performance considerably. In chapter 5, the two typical CDMA receivers are first presented: subspace RAKE receiver and minimum output energy (MOE) receiver. The detailed derivation and the simulation results of the two receivers are given. The results suggest that the performance degrades rapidly with the user increasing, because the multipath and ISI destroy the cross correlation between the user spread codes in the DS CDMA UWB systems. To improve the multiuser performance, the KALMAN filter is introduced into UWB system. We began with the characteristics of the UWB signals to get the state and observation equations, Next, based on the transverse filter structure, we obtain a more simple and practical Kalman filter, which improve the performance greatly according to the simulation results. In the following part of chapter 5, MMSE and MMSE RAKE algorithms are discussed and analyzed in detail respectively. In the final part, we compare the performance of the five algorithms above. The results show that the MMSE algorithm is the best one with large calculation complexity. MMSE RAKE is worse than MMSE and KALMAN filter is better than MMSE RAKE and worse than MMSE. But because of the calculation complexity is lower than MMSE, KALMAN filter is a good multiuser detection receiver.