Section 7.1. Introduction

7.1. Introduction

UWB signals will encounter interference from many sources, primarily from relatively narrowband (NB) systems. In addition, UWB signals will also affect a large number of NB radios. Of critical importance is the potential interference with GPS and navigation bands, as well as cellular bands. There is a rich and growing literature on UWB radios. However, issues related to interference have only been partially addressed. Here, we assess the interference caused by UWB signals via analysis and simulations. Analytical results include the aggregate effect of spatially distributed UWB radios on an NB receiver, and theoretical BER expressions. The impact of NB interference on a UWB receiver is also studied. Some results on the UWB-to-UWB interference (the multiple access issue) are presented. Simulation results are included to verify the theory.

A wideband (WB) system, by its very nature, will interfere with existing NB services in the same bands. In turn, the NB signals will act as interferers to the WB system. The extent to which performance is degraded by the interference will clearly depend upon the number and distribution of the interferers, the relative powers, and the type of modulation used. Proposed applications for UWB radios include sensor networks, Wireless Personal Area Networking in the IEEE 802.15 framework [1] for indoor and outdoor command posts, geolocation, RF tagging, LPD/LPI applications, automobile collision avoidance systems, inventory control, and so on. As such, the density of fielded UWB devices could be high. UWB is also being considered, along with other alternatives, in the 802.15 WPAN 4a alternative for low rate applications such as sensor networks [2]. The FCC specifications [3] and European standards [4] essentially limit the EIRP to US Part 15 limits, that is, 41 dbm/MHz in the range 3.1-10.6 GHz, and the UWB signal is required to occupy an instantaneous bandwidth of at least 500 MHz.^[1] There are significant differences in the Part 15 levels for intentional, unintentional, and incidental radiators. Conforming to Part 15 levels does not imply that interference to other systems, such as GPS, is harmless. The intent of the FCC mask is to enable co-existence of UWB radio services with currently licensed services, both commercial cellular as well as critical applications, such as GPS and navigation systems. A UWB radio with the pulsed Gaussian monocycle has a bandwidth of about 1.6 GHz centered around 2.0 GHz. Prototype radios with bandwidths of 3-4 GHz, centered at frequencies of 3-4 GHz have been reported. It is important to have tools to assess the potential interference with GPS and navigation bands as well as cellular bands.

^[1] Please see Chapter 1 for details.

Recent measurement campaigns indicate that GPS and some radar systems may be adversely affected, particularly if the pulse repetition frequency (PRF) is high, even though the device itself might conform to Part 15 emission limits [5-17]. With the exception of [14-17], these reports deal primarily with the impact of UWB radios on GPS.

Although there is a rich and growing literature on UWB radios [18-20, 22-24], issues related to interference have only been partially addressed. The effects of UWB interference have been considered in [25], [26], and [27], where the unconditional BER is evaluated via simulations.

There is a body of work that studies the degradation in the performance of UWB radios due to narrowband interference (NBI), such as tone jammers [28] and partial band interference (PBI) [29]. If the frequency of the tone jammer is known, several techniques can be used to shape the transmit pulse so as to avoid the interference [30-32]. In [33], spreading codes are designed that combat NBI by using frequency-spreading. The performance of generalized RAKE receivers and MMSE receivers, in the presence of MAI and NBI, has been analyzed in [34-37]. These techniques assume that all relevant user codes, multipath parameters, and noise parameters are known. Performance in the presence of multiple users has been studied. For example, [38] provides exact BER expressions for an interference limited uncoded TH-PPM and approximate expressions for the coded/asynchronous case. A Chernoff bounding approach is taken in [39], where bounds are established taking into account the episodic^[2] nature of the transmission, the variance of the additive white Gaussian noise (AWGN), and imperfect channel estimates used by a RAKE receiver. However, results on aggregate effects, that is, the impact of multiple UWB emitters on NB radios, is rather sparse: [40], [25], [8], and [41] provide both analysis and simulation results. A simulation study is also reported in [42].

^[2] In contrast with a DS-CDMA radio, in a pulsed UWB system, pulses are not transmitted continuously; instead, there is a (variable) off time between pulses. This duty-cycled nature of the UWB signal has been termed episodic in [39].

In order to develop techniques for UWB interference suppression (such as blankers and other non-linear pre-processors), it is important to develop analytical and empirical models of the interference. In this chapter, we consider the interference effects of UWB signals on typical NB radios. We also consider the effect of NB interferers on UWB radios. The analysis tools are not very different from those used in studying the impact of conventional direct-sequence spread-spectrum and other multiple access signals on classical NB systems.

The theoretical analysis takes into account the waveforms of the interferers (Gaussian monocycles for UWB, typically NB Gaussian processes for NB), the spatial distribution of the sources (density and distances), propagation losses, and receiver models. The interfering UWB signal structure will cause effects different from those due to thermal or broadband noise. In particular, the pulse repetition rate, duty cycles, burst times, and specifics of the waveforms play significant roles.

The chapter is organized as follows: In Section 7.2, we consider the impact of UWB radios on an NB radio. After reviewing the signal model, we derive an expression for the conditional BER in the presence of one or more interferers and AWGN. In Section 7.3, we provide a statistical description of the impact of many interferers. In Section 7.4, we quantify the effect of NBI on an UWB signal. Finally, in Section 7.5, we summarize some results on the UWB-on-UWB problem, that is, the (perhaps asynchronous) multiple access interference (MAI) issue. Numerous simulation results are included. Most of our analysis is at "baseband" for convenience but holds for the "passband" case with simple frequency translation.