Research On Synchronization Algorithms In UWB System

In recent years, Ultra Wideband (UWB) has been received more and more attention because of its advantages, such as high transmission rate, low power consumption, good secutity, robustness to multipath and low cost. UWB systems are very suitable for indoor or short-range high-rate wireless communications and have extensive application on accurate ranging and location. Since UWB systems transmit and receive signal by streams of very low power density and ultra-short duration pulses, even a small synchronization error could cause remarkable performance degradation which would be a serious challenge to synchronization in UWB systems. Just due to the characteristic of the UWB signal, the receiver has very high requirement for synchronization accuracy which is rather difficult to achieve. On the one hand, synchronization process needs extremely long searching time owing to ultra-short pulses (nanosecond or sub-nanosecond) and low-duty cycle (less than 10%). On the other hand, the signal with low power density and the effect of noise and interference make it difficult to reach high acquisition probability. As a result, synchronization has become a hotspot in the research of UWB.In this paper, the evolution and the fundamental principles of UWB ststems are firstly reviewed, and modulation modes, signal model and channel models of UWB systems are briefly introduced. After analyzing synchronization problem of UWB systems, many kinds of typical frame-level synchronization algorithms, such as Maximum Likelihood (ML) based synchronization algorithm, Generalized Likelihood Radio Tests (GLRT) based synchronization algorithm and Cyclostationarity based synchronization algorithm are introduced and discussed in detail. The system simulations and performance analyses are made to each algorithm. Then two kinds of pulse-level timing acquisition algorithms called Dirty Template based acquisition algorithm and Least Square based acquisition algorithm are analyzed. Aiming at their defects, a low-complexity pulse-level timing acquisition algorithm combined the advantages of them is proposed. The improved algorithm uses two-step method to obtain pulse-level acquisition. Simulation results show that the improved algorithm can achieve good performance while reducing complexity.