The recent history of Ultrawideband (UWB) communications technology has seen great debate over whether UWB causes unacceptable interference to existing users of the same and nearby bands. This article begins with the current state of regulations, followed by a discussion of the out-of-band performance of the WiMedia PHY. It then explores in depth methods for coexistence of WiMedia UWB with existing systems in the 3-5GHz band.
Since UWB covers such a large bandwidth, coexistence with incumbent users, both WiMAX and others, is important, and may in fact be required for regulatory approval in most countries outside the United States. The spectrum of a WiMedia UWB signal can be "sculpted" based on a priori decisions about restricting emissions in certain bands such as the radio astronomy bands in Japan; the OFDM tone weightings could also be derived from an environment survey done by the OFDM modem to measure the "interference temperature" of the environment. In either case, the multiband OFDM signal permits flexibility in this "spectral sculpting" that is unique to MB-OFDM UWB signals.
The original concept of UWB is that of a spectral "underlay;" in other words, the UWB signal would have such a low power spectral density that its interference potential would be negligible. The relevant FCC rules, under which UWB was authorized in a Report and Order dated February 14, 2002, state that Part 15 devices such as UWB shall not cause "harmful interference."
Unfortunately, this term is never defined. In an effort to protect incumbent services in the UWB bands while still enabling the personal area network (less than10 meter) usage model, the FCC chose to allow UWB to radiate at a power spectral density of -41.3dBm/MHz, which is the same level allowed for Class B unintentional radiationthe same power level that is considered "background noise" for consumer electronics devices. The power spectral mask for indoor UWB devices as approved by the FCC is shown in Figure 1.
As the UWB industry has undertaken the task of working with the regulatory bodies around the world to get approval outside the U.S., the debate over what constitutes harmful interference has been heated, especially in Europe. The central issue is not whether UWB can cause interference; it is possible to construct usage models of UWB that do indeed interfere with other systems. The debate is over whether these usage models are relevant and whether the interference is serious enough to warrant protection of the victim service.
There are two cases to consider: a) The incumbent signal is outside of the UWB bands, and b) The incumbent signal is in the same band as the UWB signal. The former case can be handled with careful chip, board, and filter design, while in the latter case, the interference is treated as negligible (in the FCC rules), or the UWB radio must use mitigation techniques, which include "Detect and Avoid" (DAA) and collaborative coexistence Packet Traffic Arbitration (PTA).
Interferencewhen can it occur?
Before analyzing these two cases in detail, it is useful to analyze the amount of power that can be coupled into another radio's received from a UWB transmitter. Since the FCC rules specify the UWB power as a spectral density, we merely have to examine the effective bandwidth of the incumbent receiver to determine the total power coupled into the receive chain, as long as the UWB transmitter has a constant power spectral density over that incumbent receiver's bandwidth. This can be written as an equation as follows:
Peff = the effective power available before any path loss or other losses
Beff = the effective bandwidth of the incumbent receiver
The power available to be couple into an incumbent's receiver is more complicated than this, however. As the Peff increases, so does the noise floor; the noise power that the incumbent receiver will observe in its bandwidth is, as a minimum, the thermal noise due to electron movement, which is given by:
NO = the noise power at the receiver's antenna
k = Boltzmann's Constant (8.617 339 - 10-5 electron-Volts/°K)
T = temperature in degrees Kelvin, typically taken as 300°K for room temperature
And Beff is the same as above.