As promising as it may seem, proposals combining UWB and SDR are however not so numerous. The reason for that
comes from the a priori incompatible nature of those two technologies. Indeed, SDR aims at generalizing as much as
possible the features of a communication system by digitizing the radio signal as close to the antenna as possible. Specialization of the system is consequently done in the digital domain which, thanks to its reconfiguration capabilities, can potentially adapt to any transmission scheme for the physical layer and any protocol behavior for higher layers.
This is not adequate anymore in a UWB context. The fault is that of the large bandwidth which demands very high sampling rates. Even if such conditions can be fulfilled in laboratory conditions, this is completely incompatible with products constraints in terms of performance, power consumption and cost, all the more so for embedded systems powered by capacity-limited batteries. So high sampling rates are particularly unaffordable regarding the analog to digital converters and the associated digital components required to process signals at such speed.

Necessity to reorient SDR for UWB

Given those considerations, it seems quite clear that tackling SDR in the usual manner leads to a dead-end. It is
technologically impossible to digitize bandwidths of several Gigahertz in a commercial mobile communication system
without an unacceptable extra cost. Consequently it is not consistent anymore to move the analog to digital conversion
as close to the antenna as possible to make SDR UWB. On the contrary, it seems mandatory to let the analog front-end do as much of the processing as possible so that a narrower bandwidth digitization becomes possible and sufficient, while preserving all the different potential manners of using the system. This means having some kind of SDR analog front-end.
The first condition is to have an analog front-end that delivers a signal generic enough to be further processed in different ways for the achievement of various kinds of telecommunication purposes.
The second is to realize as much of the processing as possible in a passive low-power manner, without depriving the signal from its general content. We call this analog preprocessing. The third requirement for the RF front-end is to feed
the A/D converters with signals that may be sampled at sufficiently low rates so that technological, cost and powerconsumption barriers are removed at a short or medium term.
Fig. 1 illustrates the usual SDR vision with the A/D converter as close to the antenna as possible.

In order for SDR to be viable in a UWB context, it is necessary to add an analog pre-processing stage before conversion so as to reduce the constraints put on the digital part, as shown on Fig. 2.