The release of the Ultrawideband (UWB) regulations by the Federal Communications Commission (FCC) in 2002, and the complimentary IEEE standard 802.15.4-2006 P802.15.4a amendment in 2007 for UWB Personal Area Networks (PANs), have promised high-speed wireless personal area network (PAN) communications, and have led to an increasing interest in using the 3.1 - 10 GHz radio band for transmitting large amounts of data from a body-centric antenna over distances of less than lOm. For operation in indoor environments the proximity of the human body can cause shadowing and perturbations of the transmitted signal. Radio frequency sounding is often employed to characterise the UWB radio channel in an effort to investigate such phenomena. This thesis explores statistical characterisation of the off-body S1S0 and MISO UWB indoor radio channel between 3.1 GHz and 6.0 GHz However, the sounding equipment may itself alter the radio channel parameters and to this end, a comparison of an optical and radiofrequency (RP) cable feeding the body-mounted antenna for an off-body UWB link was performed. This revealed that the RP cables add measurement error due to their reflective construction, increasing launch scattering and post-launch reflections. This was found to be particularly evident for low reflectivity environments and for non line of sight (NLOS) configurations. Indeed, the use of cables was found to increase received power, increase multipath delay and had the potential to alter the channel's statistical distribution model. This was found to be true for 89% ofNLOS and 44% of LOS experiments. The same novel optical cable fed UWB sounder was implemented to investigate off-body links in a 49 m2 hospital environment for both stationary and mobile UWB radio transmitters mounted on the waist and chest. The results showed that received signal strength values were dependent on whether transmit and receive antennas had line of sight (LOS) and were also affected by body- shadowing and antenna-body position. For mobile conditions, received signal strength tended to be Lognormally distributed with NLOS links having significantly lower mean values. Excess time delay results for the mobile user tests were best described by the Weibull distribution. Overall, the results favoured the chest mounted antenna position, with higher mean signal levels, reduced mean excess delay and less difference between LOS and NLOS channels. Additionally, it was found that the Rician k-factor values for the anechoic and hospital environment were significantly affected by environmental conditions, highlighting that the human body acts primarily as part of the radiating structure. Nonetheless, variations in the k-factor values for the differing antenna mounting positions also indicated that this was also a factor affecting off-body radiation. Finally, the performance of a multiple antenna body-centric transmitter was investigated in an office and corridor environment. It was found that when using such a multi-antenna array, all the distributions for each of the channels were Lognormally distributed in the office and Rician in the corridor. It was also found that using selection combining techniques, all the distributions for two-channel diversity and three-channel diversity for both LOS and NLOS journeys were mostly Lognormally distributed in the office and the corridor. Thus, use of simple channel diversity techniques changed the channel's statistical distribution from Rician for a single channel to Lognormal for combined channels in the corridor. A general correlation between mutual coupling between the antennas in the array and the received power values on each channel was found. It was also found that when the antenna array was placed on the body with such a configuration as presented here, diversity gains were poor, in the region of only 1 dB. Overall, the results in this thesis indicate that off-body UWB channel characterisation should be conducted with optical fed antennas and more research into antenna positioning to attain diversity gain is required.

Ultrawideband (UWB) wireless systems employ signals with bandwidths in excess of 500 MHz or with relative bandwidth more than 20%. The radiated signals have low power spectral density. A decade ago, UWB wireless systems were deemed to be the technology that will deliver 'Gigabit-wireless' for short range communications. However, the performance of current systems is significantly below the initial expectations. This thesis explores the UWB wireless channel and shows how its properties limit the performance of current UWB systems. Furthermore, it is shown that if the knowledge of the channel is fully exploited a significant performance improvement of UWB systems can be achieved. The thesis begins with exploration of the channel properties. Unlike previous work, that has investigated either the 'classical narrowband' channel with bandwidth <100 MHz or the UWB channel with bandwidth >1 GHz, this work studies the transition between the narrowband channels with bandwidth of 1 MHz to the extremely wide band channels with bandwidths of up to 10 GHz. The thesis concludes that for signals with bandwidth <1 GHz UWB antennas and antenna arrays can be described by the classical means of gain and array factor, i.e. they treat such signals as 'narrowband'. In contrast, wireless propagation for signals with bandwidth > 100 MHz has properties 'like UWB channels' with bandwidths in the GHz range. Additionally, the thesis suggests a correction to the IEEE802.15.4a model for channel impulse response because as will be shown in the thesis many multi paths in the model are manifes- tations of the antenna impulse response. Hence multiple multipaths in the IEEE802.15.4a model actually represent a single multipath component. This reduces the number of multipath components in the model by approximately factor of five. The understanding of the transition between narrowband and ultrawideband channel is used to improve the spectral efficiency of impulse radio systems which traditionally use signals with bandwidth> 1 GHz. It is shown that the optimum signal bandwidth for impulse radio systems is in the range 150-450 MHz. Such systems balance the robustness against frequency selective fading with the reduction of duty cycle. Hence, the data-rate of impulse radio systems can be significantly improved. The frequency selective fading is shown to be the main limiting factor for the performance of the commercial UWB WiMedia systems with OFDM. It is shown that adaptive loading of OFDM sub carriers , which is compatible with the frequency selectivity of the channel, is more suitable for UWB OFDM systems than the use of strong Forward-Error-Correction measures. The introduction of the adaptive OFDM is not a significant change to the design of the scheme because the commercial WiMedia standard already foresees pilot OFDM symbols for channel estimation. The adaptive OFDM for UWB has already been considered by some authors. Unlike previous works, this thesis explores the performance of such a system in a large number of measured wireless channels. Finally, the thesis studies the MIMO techniques for UWB systems. Suitable schemes for fixed and adaptive OFDM are discussed. A realistic simulation using measured wireless channel shows that a 4x 1 system with a low complexity beam-steering and adaptive OFDM can deliver a data-rate of 400 Mbps over a range of 9 m. This performance is for a system with bandwidth 528 MHz (like in the WiMedia standard). A further increase can be achieved with the increase of the system's bandwidth.