Channel ranking scheme in wireless sensor networks based on packet delivery ratio estimation

Abstract

The widespread deployment of competitive wireless technologies in the 2.4 GHz unlicensed Industrial, Scientific, Medical (ISM) band has introduced co-existence issues between different wireless devices. Co-channel interference severely affects the performance of Wireless Sensor Networks (WSNs) that are limited to low transmission power and low data rate. This problem can be mitigated by searching the candidate channels and choosing ones for operation which provide more reliable connection. In this thesis, a Packet Delivery Ratio (PDR) estimation algorithm is proposed that can be used by WSNs operating in license free bands in order to rank the channels. Since the PDR is a function of Signal to Interference plus Noise Ratio at receiver node and the traffic pattern of the interferer, receiver node predicts the achievable PDR on each channel according to received signal strength from transmitter node, noise and interference characteristics obtained through spectrum measurements. Finally, channels are ranked based on achieved estimates. The PDR estimation algorithm is implemented on the IEEE 802.15.4-based wireless sensor platform in order to assess its performance. Wireless channels are emulated to model different environments and the accuracy of PDR estimates are observed under different channel conditions. It is beneficial to minimize the energy consumption of channel scanning by reducing the number of collected samples from the channel, while the desired accuracy is fulfilled for channel ranking purpose. Hence effect of number of collected samples on performance of estimates is investigated. The channel ranking method is also evaluated in a real environment to rank IEEE 802.15.4 candidate channels in existence of interference from multiple wireless devices in 2.4 GHz frequency band.