Performance evaluation and design for communication networks of the future is conducted. Emphasis is placed on large, high-speed networks. Both optical and electronic networks are considered. Topics include (1) spectral response of queues and diffusion approximation, (2) continuous traffic in packet switches, (3) multirate circuit switches, (4) optical interconnection, and (5) dynamic load balancing. Analysis consists of a mixture of exact probabilistic and combinatorial methods and simulation. Design is motivated by modeling and analysis and aided by optimization tools of both combinatorial and nonlinear iterative types.

The goal of this research is to investigate basic issues involved in providing multimedia communications over a heterogeneous network of both wireless and wireline links. Topics under investigation include wireless multimedia networks, routing and congestion control, adaptive coded modulation for spread-spectrum communication systems, and the interface of wireless networks to high-speed wide-area wireline networks.

This research project seeks to develop robust but near-optimal methods for timing acquisition and demodulation for wireless environments. Multiuser and single-user settings and narrowband and wideband interference are considered in the design and performance evaluation of different receiver structures.

Our objective is to investigate block and lattice charged codes with a new approach and to exploit the advantages of this approach to provide bounds on decoding complexity and to develop efficient maximum-likelihood decoders. The precise trade-off between complexity and performance is studied. We also investigate modulation codes for input- constrained channels. Viewing block codes as dynamical systems makes it natural to consider applying results from algebraic coding theory for the design of modulation encoders. Ways of integrating a prescribed error-correction capability within such encoders are also studied.

We will investigate creative new techniques for reliable transmission of digital information. The main objectives are to achieve a deep theoretical understanding of the underlying problems and to develop practical coding schemes that can be implemented in real applications. Intersymbol interference channels are emphasized, as are the digital speech and image transmission channels characterized by unequal input probabilities and subjective distortion criteria. Our research comprises two major activities: to extend prior work in trellis structure and trellis decoding of block and lattice error-correcting codes and to develop novel data transmission techniques particularly suited to specific channels of practical importance and extending beyond the classical error-control approach.

This project investigates channel coding techniques for source coding applications with an emphasis on image, video, and speech coding applications. The main objectives are a theoretical understanding of combined source-channel codes and development of practical algorithms for such applications as low-bandwidth video compression and low-delay speech coding. Specifically, very narrow bandwidth transmission channels require efficient coding schemes to protect the transmitted source information from channel error corruption. The project also explores low-complexity techniques needed for low-delay real-time implementations.

Long codes and high-dimensional constellations are necessary to achieve high coding gains over the uncoded quadrature amplitude modulation signaling. Currently, the complexity of decoding high-dimensional constellations is well beyond the reach of today's technology. In this research, we are developing new techniques for efficient bounded-distance decoding of high-dimensional signal constellations. We anticipate that these techniques will make coding with such constellations not only feasible in principle but practically implementable with high-speed, low-power hardware, which in turn will make it possible to achieve very high effective coding gains on band-limited Gaussian channels, at an affordable complexity.

This research addresses the development of single-user and multiuser communication techniques that lie at the intersection of communication theory, information theory, and error control codes and studies the effects of amplitude constraints on the capacity of and signaling for practical channels. The second activity investigates the practical use of error-control encoding/decoding schemes matched to certain transmission and storage channels. The third area investigates source coding for packet networks and narrowband channels. We examine the effects of channel errors from theoretical and algorithmic perspectives in order to develop effective data compression and coding techniques for such applications as speech and video in power-constrained environments.