Mathematics Colloquia and Seminars

Return to Colloquia & Seminar listing

Time reversal for wireless communications

Applied Math

Speaker: Persefoni Kyritsi, Stanford University
Location: 693 Kerr
Start time: Fri, Apr 23 2004, 4:10PM

Time reversal (TR) is a powerful technique to achieve temporal and spatial focusing. In this talk, we investigate the temporal focusing potential of TR in the context of fixed wireless access (FWA) communications. The measure of comparison is the ‘delay spread’ of the channel impulse response. The efficiency of TR depends strongly on the available bandwidth, the nature of the propagation channel and the number of transducers. Specifically for wireless applications, spectrum is the most expensive commodity and therefore bandwidth is limited, the radio channel has directional characteristics and varies with time. Moreover, wireless devices have limited capabilities (antenna directionality, power amplifier limitations). All these factors affect the performance of the TR technique. As a basis of our analysis, we present experimental data taken in an urban environment over a bandwidth of 20MHz at 5GHz, for various transmitter-receiver separations and under various scattering situations. We first elaborate on the power considerations that the system constraints introduce and discuss optimal and suboptimal power allocation strategies. Using the actual channel measurements, we demonstrate that TR is a powerful way to decrease the perceived delay spread of the channel in the scenarios under investigation. The application of TR requires knowledge of the channel transfer function at the transmitter location. This can be achieved with the use of channel state information (CSI) feedback (if the intended receiver itself estimates the channel), or with knowledge obtained in the uplink. The efficiency of TR depends on the accuracy of this knowledge, and, obviously, more advanced power allocation schemes suffer more from outdated channel state information. We illustrate this effect by calculating the degradation in delay spread reduction introduced by delayed CSI feedback. We show that with perfect CSI, the delay spread of the channel can be reduced by up to a factor of 3, and for reasonable feedback delays, the benefit of employing TR remains significant. Moreover we show that the performance of a minimum mean square error equalizer with a finite complexity improves significantly after the application of the time reversal technique.