Communication through high delay spread X bandwidth (HDB) channels

There is increasing demand for higher throughput and reliability in wireless communication systems. A number of techniques have been used to address such needs. These include high channel bandwidths (wideband), use of multiple antennas (MIMO), adaptive modulation and turbo coding. Typical examples of such wideband systems include IEEE 802.15.3 Ultra Wideband, IEEE 802.11 Wireless LAN systems, and IEEE 802.16e Wireless MAN systems.; In these wideband systems, the product of channel delay spread into bandwidth is usually large. This means that the impulse response in such systems has a large number of resolvable multipaths. Such rich multipath channels also occur in ultrasonics and under water acoustics. Over the past ten years, a concept called Time reversal (TR) has been developed in these domains. TR uses a prefilter that is the time reverse of the channel response. TR results in three striking effects - spatial focusing, temporal focusing and channel hardening. Spatial focusing means that the retransmitted signal arrives at the target antenna at a higher power level than the surrounding medium. Temporal focusing means that the retransmitted signal is recompressed into a narrow pulse at the target antenna. Channel hardening refers to significantly reduced signal fading or variability. In a sense, TR attempts to reverse the channel effect in the forward path - energy is dispersed in space, delay and the individual paths exhibit fading; TR partially restores the property of the original pulse which was highly localized in space and time, with a fixed amplitude.; In this thesis we explore the value of TR in wireless systems that use high delay spread channels. We show experimental evidence of TR in wireless. We then go on to study effect of TR on channel capacity, its value in equalization, and also explore certain generalization of TR to improve energy localization. We show that while TR is not optimal from a capacity or equalization viewpoint, a generalized TR shows more promise in spatial and temporal equalization. We show that a certain tradeoff called rate back-off is needed to realize spatial and temporal equalization. We also study the statistics of spatiotemporal focusing.