Channel Selection in Spectrum Agile and Cognitive MAC Protocols for Wireless Sensor Networks

Wireless Sensor Networks have emerged as a promising technology for enabling embedded sensing in many daily life applications. The ever growing popularity of wireless devices has caused crowding of the existing spectrum. One of the major issues faced by real sensor network deployments is the coexistence with other devices and symbiotic interoperability in uncoordinated wireless environments. Since sensor networks consist of low power embedded nodes, the issue of coexistence and interference avoidance becomes important as the nodes remain handicapped while competing against other less resource constrained devices for the same channel. Surprisingly only a limited amount of work has been done towards designing spectrum agile and cognitive MACs catering the inevitable wireless interference. The selection of an interference minimal channel for data communication is key for reliable operation of a sensor network MAC protocol. In this paper, we identify the fundamental design challenges for dynamic selection of an interference free channel in wireless sensor networks. We describe a lightweight channel selection strategy designed to efficiently cater the real world deployment issues. We have carried out performance evaluation of our scheme on real sensor nodes and measured metrics such as successful packet delivery ratios and throughput. We provide comparative analysis of our channel selection algorithm with a random channel scheme on a sensor node testbed in the presence of different types of wireless interference patterns. Experiments show that our scheme has fast convergence and is able to achieve higher throughput and successful packet delivery ratios.