The ``wireless revolution"" and the ever increasing need to communicate on the move is continually bringing up new engineering problems and challenges. One of these challenges is the design of communication systems capable of delivering high data rates over random time-varying channels. An extensive body of work exists on the analysis and design of mobile wireless communication systems. Almost invariably these works assume that the channel variations are relatively slow so that the channel can be considered to be approximately constant over a block of symbols. Thus, the fundamental operation of existing mobile communication systems is the same as in systems designed for time-invariant channels. Various methods such as powerful error correcting codes, channel equalizers, and spatial or temporal diversity are used post-facto to mitigate the effects of the time-varying channel. This research project is developing novel communication systems for channels which vary so rapidly that the performance of conventional communication systems is severely degraded.
A fundamental assumption underlying the design of a wide range of digital communication systems is that the receiver has the ability to estimate the radio or acoustic propagation channel, and to use that estimate in the detection process. In random rapidly time-varying channels with large Doppler and delay spreads, reliable channel estimation is not possible. This project considers the problem of designing a communication system when the channel realization is unknown and inestimable, and only the statistics of the channel are known (or can be estimated). Thus, the channel realization is only known to lie somewhere in a subspace whose dimension is the rank of the channel covariance matrix. The unit-rank case reduces this to a known class of communication problems. However, the case where the rank is greater than unity requires a significant extension of current theory. This problem formulation leads to new receiver structures and novel ways of designing communication systems which must operate in harsh time-varying environments. This project is developing optimal and sub-optimal solutions for communications through subspace channels and an analysis of their performance. Various types of communication systems are being studied including: single-input single-output systems operating through frequency selective channels, MIMO systems, and multi-user and multi-access systems.
