Featured, Publications

Modeling protocols in relay networks with Markov chains

Different communication protocols for relay selection in wireless networks have been modeled using Markov chains in order to analyze and compare their performance.

Written by Christian Bettstetter

Wireless communications often suffers from error bursts — a long sequence of bits being corrupted when being transmitted over the air. A relatively new technique that tries to mitigate this unreliability is cooperative relaying. Its basic idea is simple: When a device transmits data packets to a destination, adjacent devices can overhear these packets. If the direct transmission to the destination fails, one of those devices can retransmit (“relay”) its packet copy to the destination. Information theorists have proved that such relaying can outperform standard communication techniques. In practice, however, relaying requires some coordination among the involved devices. In particular, a protocol for relay selection is used which in turn causes extra signaling packets. Such overhead can reduce the throughput benefits gained by relaying, and a relay update policy must define how and when relay selection is performed anew.

Nikolaj Marchenko and I investigated the performance of cooperative relaying with relay selection. Our forthcoming article to be published in the IEEE Transactions on Vehicular Technology proposes a framework for modeling and analysis of cooperative relaying protocols with the help of semi-Markov processes. Using this tool, we find under which conditions which protocol leads to improved throughput and energy efficiency.

“Reactive selection is commonly considered to be the best,” Marchenko explains. “Though, if you take signaling and energy for receiving into account, it turns out that reactive selection can be very inefficient. Under certain conditions it is outperformed even by a single permanent relay.” The PhD candidate and I propose a novel technique that adjusts to the current situation in an adaptive manner. After having laid the theoretical foundation, my team is now working on an experimental evaluation of cooperative relaying in industrial sensor networks.

Publication

Nikolaj Marchenko and Christian Bettstetter. Cooperative ARQ with Relay Selection: An Analytical Framework Using Semi-Markov Processes. IEEE Transactions on Vehicular Technology, vol. 63, no. 1, pp. 178–190, January 2014.