The performance of impulse radio is investigated for ultra-wideband (UWB) channels and orthogonal M-ary pulse position modulation (PPM) in terms of the maximum achievable information rate, where the outputs of the M-ary correlation receiver implemented through the rake structure with conventional maximal ratio combining are subject to either soft or hard decisions. The major difference observed between the soft- and hard-output systems is that at low values of the mutual information the minimum information bit energy to noise ratio required for reliable communication increases with the constellation size for the hard-output systems. Furthermore, it is demonstrated that independent of the type of the output, achievable rates are considerably low at medium-to-large signal-to-noise ratio (SNR) values in UWB channels compared to those in additive white Gaussian noise channels, as the frequency-selectivity of the UWB channel destroys the orthogonality of the transmitted PPM signals. Interference from other users and constellation size put additional constraints on the achievable rates at high SNR.

In this paper, the performance of a time-division duplex ultra wideband system that has a transmitter/receiver pair of rake combining structures is optimized, where the total number of rake fingers to be deployed at the transmitter and the receiver is fixed. It is shown that there exists an optimum distribution of fingers between the two structures for systems with intersymbol interference, which maximizes the signal-to-interference plus noise ratio. Depending on the total number of rake fingers and/or post-rake fingers, i.e., those at the receiver, the optimum placement of post-rake fingers changes as the simulation results demonstrate.

The performance of impulse radio is investigated for frequency-selective ultra-wideband (UWB) channels and orthogonal M-ary pulse position modulation (PPM) in terms of the maximum achievable data rate, where the outputs of the M-ary correlation receiver implemented through the rake structure with conventional maximal ratio combining undergo either hard or soft decisions. The major difference observed between the soft and hard decision output systems is that at low code rates increasing the constellation size, M, requires an increase in the minimum signal-to-noise ratio (SNR) per communicated bit for the system with hard decisions. It is demonstrated that independent of the type of decision making, achievable rates in UWB channels are low compared to those in additive white Gaussian noise channels, as the frequency-selectivity of the UWB channel destroys the orthogonality of the transmitted PPM signals. The interference from the other users and the constellation size put additional constraints on the achievable data rates at high values of SNR.

The frequency-selective ultra-wideband (UWB) channels are subject to various types of electromagnetic noise, which is non-Gaussian (impulsive). The conventional linear receivers based on the additive white Gaussian noise assumption exhibit performance degradation. In this paper, a robust receiver using the M-estimation technique is introduced for multiple access UWB channels with impulsive noise, eliminating multiuser interference and impulsive noise at the same time, provided that accurate channel information about the users is available. Both the robust receiver and its simplified version are shown to outperform the linear receiver based on the Gaussian noise assumption.