Ultra Wide Band (UWB) systems represent a promising approach allowing the development of new types of wireless Local Area Networks and Sensor Networks. For twenty years, UWB has been an active field of research and it could be the physical layer for future multiple access wireless networks with very high bit rates, as well as low bit rate low power consumption networks with high accuracy positioning abilities. We address in this thesis the issues related to UWB ad-hoc sensor networks from the physical layer towards the upper layers: antenna optimization issues, interference modelling, IR-UWB Advanced Receivers and Transmitters with efficient wireless communication, throughput optimization, and cooperation issues. A first contribution deals with the IR-UWB non-Gaussianity of the MUI distribution. We have developed a performance criterion that takes into account the system's parameters and the MUI distribution in order to compute the capacity of the communication and its throughput. We have also shown that throughput can be maximized by controlling the impulsiveness of the interference, and we have presented the impact of antenna design (and pulse shape) on the impulsiveness and so on the capacity of an IR-UWB system. Taking into account the study on the MUI, another contribution is concerned with novel IR-UWB receivers and transmitters. The proposed receiver assumes that the Probability Density Function (PDF) describing the statistic of the interference is a mixture-based distribution. We have also investigated a novel transmitter design based on Time Reversal (TR), and we have studied the gain brought by time reversal when the complexity is switched from the receiver to the transmitter i.e., when the number of fingers is increased in the prefilter while it is reduced in the Rake receiver. Moreover, we showed that TR combined with Rake technique changes the PDF of the MUI. The work has further focused on the cooperation for wireless networks. A first study proposes an incremental relaying scheme that uses multiple relays, but we have also investigated a practical incremental relaying scheme that deals with the imperfect feedback, and finally a simple distributed relay selection policy without any need of information exchange between the relays.
