AN experimental methodology for evaluation of multi-hop medium access control protocols of wireless sensor networks

Abstract

Since wireless sensor nodes are battery powered, the emphasis was on reducing the energy consumption of the sensor nodes. However, with the increase in real-time traffic requirements in wireless sensor nodes, network performance metrics are increasingly important. Nonetheless, most existing network performance studies were done via simulation, with few experimental studies performed on testbeds using ad hoc methodologies. This paper proposes a structured experimental methodology for evaluating the delay, throughput, and packet loss performance of wireless sensor networks, while addressing practical issues related to the use of an experimental approach. A global clock synchronization technique, in addition to insertion of timestamps in the network packets was proposed for more accurate measurement of the network delay. Several different physical topologies such as single cluster, chain, and cluster tree topologies were defined as standard test topologies. The EMLMAC protocol was used to demonstrate the viability of the proposed structured experimental methodology. Based on the proposed experimental approach, the measured throughput for EMLMAC for the single cluster two-hop and three-hop chain topologies differed by 1.02%, 6.75% and 13.91% respectively from the simulation results. This indicates that an experimental approach is essential for verifying the actual performance of a wireless sensor network.