The routing algorithms for Wireless Sensor Networks (WSN), such as RPL, have been designed around a limitation, energy.

Hitherto, they have ignored a main strength of WSN - the ability to sense the environment that they operate on.

We see an opportunity to take better routing decisions on WSN by harnessing the onboard sensors on the sensor motes.

As an specific case of this general idea, we are currently investigating how temperature sensors can be used to improve the routing protocol RPL.

We have already established that the node-temperatures affect the link quality. The transmission power and the received signal strength diminish with the increase in the node-temperature. As much as 8dB reduction in the RSSI has been reported for an increase of the temperature by 40C. Such diurnal temperature fluctuations are observed in outdoor sensor network deployments and a mere 5dB loss of RSSI is sufficient to render a link with 100% PRR useless.

Routing algorithms must respond to the changes in link quality. We have evidence to show that the RPL topology can change drastically under temperature fluctuations thus causing data losses. In this case, we know the cause of the link quality change - temperature. We have models to relate the change in the transmission power and RSSI to the temperature changes. However, the current implementations of RPL do not have a mechanism to take advantage of the temperature measurements to respond to link quality changes in advance before they cause data losses.

Currently, we are in the process of enhancing RPL (ContikiRPL) with temperature hints to make it robust against temperature fluctuations.