The protocols and architectures of today's ad hoc networks are derivatives of those developed in the past for the Internet, which in turn date back to the ARPANET design in the 1960s. This is the wrong approach for mobile ad hoc networks (MANET), because they are very different than wired networks. In particular, many-to-many communication is intrinsic to MANETs, because of their decentralized peer-to-peer nature and the broadcast nature of wireless channels. In contrast, the protocol stacks used today are designed to avoid multiple access interference (MAI) to support one-to-one communication.
This project proposes a blank-slate approach to the modeling and design of communication protocols for ad hoc networks and MANETs. The project will seek improvements on the order capacity of ad hoc networks by means of architectures and protocols for many-to-many communication enabled by the statistical multiplexing of wireless spectrum, processing, and storage resources in such networks. The project will re-examine commonly held notions about the way in which channel access and information transport should be accomplished when in-network storage and processing are affordable, thus making multi-packet reception (MPR) a viable approach. This project will provide fundamental advances to the understanding of (a) approaches that enable many-to-many communication in MANETs by exploiting processing and storage complexity in mobile nodes; (b) how to model communication protocols operating in MANETs; and (c) what are more meaningful fundamental limits for the dissemination of information over MANETs when many-to-many communication is allowed.
