Introduction

Communication receivers have generally been designed and analyzed under the assumption that the additive noise in the system is Gaussian. Additive Gaussian noise is an appropriate model for thermal noise. Radio frequency interference (RFI), however, is well modeled using non-Gaussian impulsive statistics and can severely degrade the communication performance of receivers designed under the assumption of additive Gaussian noise. RFI is a significant problem for terrestrial wireless communication systems (e.g. cellular, WiFi and Wimax) and wired communication systems (e.g. DSL, Ethernet and powerline).

RFI can come from

  1. users and services in the same frequency band (co-channel interference)
  2. leakage from users and services in adjacent bands (adjacent channel interference) and
  3. non-communication sources.
Co-channel interference comes from dense spatial reuse of available radio spectrum in terrestrial wireless communications and from in-bundle "alien" crosstalk in wired DSL and Ethernet systems. In terrestrial wireless communications, WiMax transmission in the 2.5 GHz band leaks into the adjacent 2.4 GHz WiFi band. Some laptop manufacturers disable WiFi when WiMax is on, and vice-versa. Non-communication sources of interference include the computational platform (e.g. laptop or handset), microwave ovens (in the case of the 2.4 GHz band) and switching electronics (in the case of powerline communication systems).

Our research will enable engineers to design interference-aware communication systems to achieve 10-100x reduction in bit error rate, or 2x improvement in bit rate. The specific goals of this research are to develop

  1. statistical models for aggregate RFI at the receiver
  2. parameter estimation methods for the statistical models
  3. parametric and non-parametric RFI cancellation techniques
  4. RFI modeling and mitigation toolbox in MATLAB


Mail comments about this page to bevans@ece.utexas.edu.