Turkish Journal of Electrical Engineering and Computer Sciences




In battery-operated wireless sensor networks, the quality of service requirements such as throughput and energy efficiency must be stringently maintained while the least amount of energy is consumed. Since a major portion of energy is spent by the transceiver operations, transmission power control (TPC) in the medium access control (MAC) layer can bring about considerable energy efficiency. Since TPC algorithms will have a direct impact on the received signal strength index (RSSI) at the receiver, RSSI is the primary input parameter for any TPC algorithm. The objective of the proposed work is to decide on the exact value of transmission power required for the next transmission that will ensure an RSSI just above the threshold level at the receiver. Since this involves estimation of RSSI for the next transmission, we propose three state estimation techniques, namely Kalman filter (KF), extended KF (EKF), and unscented KF (UKF) to predict the RSSI accurately. This predicted value is used as an input for an artificial neural network (ANN)-based TPC algorithm. The effectiveness of the estimation techniques is verified by the prediction error. The accuracy of prediction is reflected in the TPC algorithm in terms of reduced power utilization.


Wireless sensor networks, RSSI prediction, state estimation techniques, transmission power control, Kalman filter, neural network controller

First Page


Last Page