With the ever growing demand for ubiquitous connectivity and data access, there is a driving need to deploy ever more extensive and higher capacity network infrastructure---wired and wireless. With such increases in network density, the propagation environments become increasingly challenging, and new communication techniques, architectures, and protocols are needed to meet these critical challenges. Examples include intersymbol (or self-), inter-channel, intra-network, and extra-network interference. Over the years, the error-control coding community has developed a wide range of codes for efficiently mitigating the effects of noise in communication systems. Ultimately, the goal of this research can be viewed as developing a signaling architecture that efficiently and effectively transforms interference into standard, more benign noise from the perspective of the underlying code.
To approach these challenges, this research develops the role of super-Nyquist signaling formats in modern coded digital communication systems. In traditional systems, the symbol rate is chosen to match the channel bandwidth. With this classical approach, the transmit pulses can be designed to be orthogonal, corresponding to signaling on independent degrees of freedom. However, in systems with super-Nyquist signaling, the symbol rate is chosen to be significantly higher than the channel bandwidth, resulting in a transmission with self-interference, whose effects can be compensated through the use of appropriate equalization. The investigation develops the role of super-Nyquist coding in a range of network scenarios, starting from simple point-to-point intersymbol-interference channel models, and progressing to richer multi-input/multi-output, multiple-access, and interference channel models. The research emphasizes the special role that such signaling plays in joint design of the physical and link layers in the protocol stack. Dual problems in source coding are also explored.
