Assessment of proactive transmission power control for wireless sensor networks

In order to prolong lifetime of Wireless Sensor Networks (WSN), Transmission Power Control (TPC) techniques are employed. The existing TPC schemes adjust the transmission power mostly reacting to changes at link quality between communicating nodes. Proactive TPC has been proposed in the recent past as reactivity does not address the need for reliability. Efficiency of a proactive TPC is determined by its prediction accuracy to link quality, ease of configuration and energy efficiency. Current state-of-the-art does not argue about proactive TPC methods on those requirements. This paper provides a targeted analysis of four prominent algorithms such as Discrete Kalman Filter (DKF), Exponentially Weighted Moving Average (EWMA), Simple Moving Average (SMA), Weighted Moving Average (WMA), and Linear Regression (LR) that could be employed in a proactive low-power TPC technique. Our experiments indicate that prediction accuracy of DKF has the least forecasting error and outperforms the prediction accuracy of all other algorithms under discussion. Amongst the Moving Average algorithms, the prediction accuracy of WMA is significantly better and linear regression algorithm has the worst performance. Evaluating the cost involved in terms of radio power and ease of configuration, WMA is the best algorithm for implementing proactive TPC.