本文提出一种采用循环前缀和脉冲调制的超宽带无线通信系统（UWB-CP/PM），以及相应的信号处理算法。借鉴采用循环前缀的单载波通信系统（SC-CP）的基本思想，UWB-CP/PM系统通过变换域均衡可以很好地抑制多径干扰。借助脉冲调制，UWB-CP/PM系统可以大大简化收发器件。利用脉冲重复周期大于脉冲宽度的特点，UWB-CP/PM系统不仅可以通过多支路接收和并行处理降低系统对基带处理速度的要求，还可以通过变换域信号处理有效地抑制衰落、多用户干扰和窄带干扰。仿真结果验证了上述性能。

在研究单次频偏累积相偏估计的基础上，本文详细分析了两次独立的频偏累积相偏估计带来的频偏捕获范围的提高。并在此基础上提出一种基于训练序列的载波同步算法。这种算法利用时域两次独立的不同分辨力的相偏估计，解决了单次估计可能存在的相位模糊问题，从而实现了大频偏载波同步。通过理论分析给出了这种算法的性能，并通过仿真进行了验证。由于完全在时域进行估计，和其他大频偏载波同步算法相比，这种算法计算量小，且实现简单

A multi-symbol detection (MSD) method based on the approximation of ML decoding for detecting the signal modulated by PPM in unknown UWB channels is proposed. The explicit expression of the achievable rates of PPM with MSD is derived, from which we study the relationship between the achievable rates, SNR and channel bandwidth. The numerical results show that multi-symbol detection significantly improves the achievable rates over conventional non-coherent detection. When channel bandwidth tends to infinity, as long as symbol number N keeps up with the resolvable multipath number L, the achievable rates will tend to AWGN capacity.

RAKE接收机常被超宽带（UWB）系统用来收集信号的能量和减小信号衰落。本文从缩短信道时延扩展的角度出发，研究RAKE接收机的部分信道均衡能力。本文推导出了RAKE接收机均衡性能与RAKE接收机的复杂度和信道衰减系数之间的关系；分析了信道估计误差对RAKE接收机合并前后多径信道变化的影响。研究表明当存在估计误差时，RAKE接收机仍然能实现缩短信道的作用，但是性能有所下降；由于低复杂度RAKE接收机引入的估计误差较少，在低信噪比条件下其缩短信道性能甚至会优于高复杂度的RAKE接收机。

The iterative algorithm for minimum variance distortion- less response filter based on auxiliary-vector (IMVDR-AV) is generalized in this paper from different perspective. By extending the optimization criteria on filter design, it is generalized to an iterative algorithm of linear constrained minimum power filter (ILCMP-AV). Starting from this algorithm, we present an iterative algorithm of LCMP filter based on local optimization criterion (ILCMP-LOC) which converges rapidly by further generalizing the conditional optimization criterion on weighting coefficient computation. For any positive definite input auto- correlation matrix and any linear constraint, the ILCMP-AV algorithm recursively generates a sequence of auxiliary vectors by maximizing the magnitude cross correlation under some constraint conditions and its corresponding weighting coefficients by minimizing the filter output variance. Theoretical analysis illustrates that the combination of the updated filters and the weighted auxiliary vectors forms a sequence of filters that converges to the LCMP solution.