Signal Processing and Speech Communication Laboratory
homephd theses › Modeling and Mitigation of Narrowband Interference for Non-Coherent UWB Systems

Modeling and Mitigation of Narrowband Interference for Non-Coherent UWB Systems

Status
Finished
Student
~Yohannes Alemseged Demessie
Mentor
Gernot Kubin
Research Areas

A transmitted-reference (TR) impulse radio (IR)-ultrawideband (UWB) system, in conjunction with an autocorrelation receiver (AcR), collects all energy of the multipath channel response at low complexity, providing an alternative to higly complex coherent reception. The analog AcR front-end, however, is open to any interfering signal in the UWB spectral regime, which undermines the inherent immunity of UWB systems against narrowband interference (NBI). The goal of this thesis is to investigate the impact of NBI and devise a mechanism against the NBI, so as to benefit from the full potential of the TR-UWB and IR-AcR combination. First, a generic data model for TR-UWB systems employing multichannel AcR front-ends is introduced. An elaborate statistical analysis of the data model components is done to characterize the signal-to-interference ratio (SIR) at the correlation stage. Based on the insight obtained from the analysis, different linear post-processing schemes are investigated. The theoretical performance limit of the linear NBI mitigation method is addressed, as a function of the number of correlators. To extend the NBI suppression capability of the post-processing stage, a front-end NBI filter is proposed, followed by the linear combining stage. Finally the study is extended to another non-coherent sub-optimal receiver, the energy detector (ED). Commonly, this receiver has a single square-law device followed by an integrate-and-dump unit, thereby experiencing a similar NBI problem as the AcR. A new receiver, the multichannel energy detector (MED), is introduced where the energy detection unit is assisted by auxiliary channels to capture the interference. Then NBI suppression is enabled by applying the linear NBI mitigation algorithms developed for the AcR. Performance evaluations are presented for the proposed mitigation algorithms, based on computer simulations. An IEEE802.11a Wireless Local Area Network (WLAN) signal with a center frequency in the vicinity of 5GHz has been assumed as a representative NBI signal. In an interference limited environment, performance gains of up to 48 dB of SIR have been shown, employing an eight-channel AcR with linear NBI mitigation.