Bilkent University
Department of Computer Engineering
S E M I N A R

 

Bandwidth-Aware and Energy-Efficient Stream Multicasting Protocols for Wireless Multimedia Sensor Networks

 

Burcu Yargicoglu
MSc. Student
Computer Engineering Department
Bilkent University

In recent years, the interest in wireless sensor networks has grown and resulted in the integration of low-power wireless technologies with cameras and microphones enabling video and audio transport through a sensor network besides transporting low-rate environmental measurement-data. These sensor networks are called wireless multimedia sensor networks (WMSN) and are still constrained in terms of battery, memory and achievable data rate. Hence, delivering multimedia content in such an environment has become a new research challenge. Depending on the application, content may need to be delivered to a single destination (unicast) or multiple destinations (multicast). In this work, we consider the problem of efficiently and effectively delivering a multimedia stream to multiple destinations, i.e. the multimedia multicasting problem, in wireless sensor networks. Existing multicasting solutions for wireless sensor networks provide energy efficiency for low-bandwidth and delay-tolerant data. The aim of this work is to provide a framework that will enable multicasting of relatively high-rate and long-durational multimedia streams while trying to meet the desired quality-of-service requirements. To provide the desired bandwith to a multicast stream, our framework tries to discover, select and use multicasting paths that go through uncongested nodes and in this way have enough bandwidth, while also considering energy efficiency in the sensor network. As part of our framework, we propose a multicasting scheme, with both a centralized and distributed version, that can form energy-efficient multicast trees with enough bandwidth. We evaluated the performance of our proposed scheme via simulations and observed that our scheme can effectively construct such multicast trees.

 

DATE: 31 August, 2010, Tuesday @ 10:30
PLACE: EA 409