Coping with unreliable channels: Efficient link estimation for low-power wireless sensor networks

The dynamic nature of wireless communication and the stringent energy constraints are major challenges for the design of low-power wireless sensor network applications. The link quality of a wireless link is known for its great variability, dependent on the distance between nodes, the antennapsilas radiation characteristic, multipath, diffraction, scattering and many more. Especially for indoor and urban deployments, there are numerous factors impacting the wireless channel. In an extensive experimental study contained in the first part of this paper, we show the magnitude of this problem for current Wireless Sensor Networks (WSNs) and that based on the overall connectivity graph of a typical multihop WSN, a large portion of the links actually exhibit very poor characteristics. We present a pattern based estimation technique that allows assessing the quality of a link at startup and as a result to construct an optimal neighbor table right at the beginning using a minimum of resources only. Our estimation technique is superior compared to other approaches where protocols continue to decide on the fly which links to use expending valuable energy both for unnecessary retransmissions and recursive link estimation.