King Fahd University of Petroleum & Minerals

  Computer Engineering Department

     Research Assistant

        The IEEE 802.11 based wireless local area networks (WLANs) have enjoyed popularity in many areas. Broadcasting in IEEE 802.11 does not support any MAC layer recovery as a result due to interference and collisions in the wireless networks, this mechanism leaves broadcasting basically unreliable. In highly dynamic environments like Rescue missions or RoboCup the communications between multiple robots is performed by broadcasting the data to its team mates called an Auction, and the broadcasting of the data in these environments should be fast and Reliable. In these environments each Robot acts like a Peer. Hence there is a need for an Ad-Hoc Wireless Networking Communication Model using IEEE 802.11 wireless interface that should be fast, reliable for the effective implementation of dynamic cooperative architectures in Multi Robot Systems. The currently available techniques are based on client-server and Publish/Subscribe communication model which are not suitable for the above problem. In this research we solved the problem of reliable broadcast over IEEE 802.11 by employing some schemes on top of application layer of IEEE 802.11b to achieve Reliability for broadcasting the data. We proposed and design a Reliable Peer to Peer Communication Protocol by exploring a number of algorithms for an Ad-Hoc network employing IEEE 802.11b wireless interface, implement those algorithms on eight Stargate board forming an Ad-Hoc wireless network and evaluated their performance such as time delays, reliability and power requirements.

 Seminar

        Parallel computing has emerged as an indispensable tool for problem solving in many scientific domains during the course of the past 20 years as parallelism is the only avenue to achieve the required performance, which is vastly higher than what is possible with single processor systems only. we explored the e1350 IBM eServer cluster of information technology center of KFUPM.we generated parallel code for a specific set of processors based on sequential program written in C-like such as the Alternating Direction Integration (ADI), the Red-Black Relaxation loop, and Matrix Multiply, etc., Collected results (like execution time of parallelized programs) and evaluated performance of e1350 IBM eServer cluster and used parallel compiler.

Disclaimer: Material are provided on this page for the sole purpose of timely information dissemination, and the copyrights belong to the authors and the corresponding publishers unless otherwise stated.

   Copyright © 2008