Study On The Performance Of Autocorrelation Receiver For TR-UWB Systems

The combination of wireless communication and computer network brings out the wireless network, which has the advantages of instantaneousness, convenience and high data rate. It is suitable for short-range and high data rate wireless communications. Among different technologies of wireless network, the ultra-wideband (UWB) has a wide bandwidth up to GHz and is able to provide a high data rate up to Mbps, thus it become a hot spot in current researches. However, there are some key problems must be solved before it is matured, one of which is the receiver design. It requires not only a good performance, but also a simple structure.Wideband wireless communication systems usually recover the transmitted signals by using a RAKE receiver. And the LGR-UWB (Locally Generated Reference UWB) systems just adopt the RAKE receiver to combine the plenty of multipath signals to improve the system performance. But the RAKE receiver needs a channel estimation and synchronization, which results in a complicated structure, and this is not what we want. TR-UWB (Transmitted-reference UWB) systems adopt the autocorrelation receiver (AcR) to carry out the differential detection. The AcR does not need a channel estimation and synchronization, so it has a low complexity, but the performance of AcR is inferior to the RAKE receiver.The main factor, affecting the performance of the AcR, is the noise contained in the received reference signals. However, the noise is white and has a zero mean, which provides a way to suppress it. That is, by arithmetical averaging, the noise can be effectively suppressed. Under some assumptions, the existed researches average the received reference signals to improve the performance. While this thesis improves on the traditional AcR to make it fit the real instance without any assumptions. The improvement averages the received multipath signals from the same bit to suppress the noise in the received reference signals and data signals respectively, and then correlates the averaged signals to obtain the decision metric. Since the reference signals and data signals come from the same bit, they suffer the same channel and have a better correlation, which has a positive effect on the system performance. As the channel delay increases, the multipath fading becomes more and more serious, hence, we combine onlythe first few arrived multipath signals to improve the efficiency of the receiver without sacrificing the system performance. As to the performance analysis, we adopt the Gaussian approximation method to deduce the mean and variance of the decision metric to obtain the BER formulas, and then prove it by simulation results. Finally, we draw the conclusion that the improved AcR has a much better performance compared to the traditional AcR.