EAGER: One-off/Continuous Convergecast and Broadcast Scheduling in Probabilistic Wireless Mesh Networks

Georgia State University Research Foundation, Inc., Atlanta GA

Investigators

Abstract

Packet transmission is one of the primary operations conducted by Wireless Mesh Networks (WMNs). All the existing work for packet transmission in WMNs is based on the Deterministic Network Model (DNM), where a pair of nodes is assumed to be either connected or disconnected. However, this assumption is not always practical and reasonable. Mesh clients are usually connected wirelessly to mesh routers resulting in unstable links. This project seeks a new exploratory study on one-off/continuous convergecast and broadcast scheduling under a more realistic Probabilistic Network Model (PNM), where a certain communication success probability is assigned to each link. Factors, such as different MAC protocols, pipeline technologies, and multi-radio multi-channel, are also taken into account for a comprehensive design and theoretical capacity analysis. For convergecast, the following four aspects are investigated: network model definition, network partition methods to identify non-inference zones, scheduling techniques for concurrent transmissions, and capacity analysis. For broadcast, Connected Dominating Set based schemes are developed. The efficiency of the proposed algorithms and the pipeline techniques are evaluated theoretically and through simulations and experiments. The success of this project provides a general framework for packet transmission in probabilistic WMNs with unstable links. Furthermore, this project also has great impacts on other kinds of wireless networks which similarly demonstrate the probabilistic characteristic. The research results will be integrated into education, and widely disseminated through conference/journal publications.