Skip to Main Content
Maximum-likelihood (ML) and generalized likelihood ratio test (GLRT)-based data detection schemes for an ultra-wideband communication system are proposed and compared with the training-based (TB) transmitted reference (TR) receiver. The exact probability of error for the TB receiver and a lower bound for the ML and the GLRT receiver are provided in terms of a doubly noncentral F-variate. An optimization with respect to the integration interval length and the energy allocation between the training and the data symbols is investigated. Analytical and simulation results reveal that the proposed detection schemes provide significant performance improvements in terms of bit-error rate over the TB receiver structure.