A new fuzzy logic approach for reliable communications in wireless underground sensor networks

Nowadays, the exploitation of Wireless Underground Sensor Networks (WUSNs) remains challenging because of the attenuation of wireless underground communications. This issue widely affects the reliability of communications in such network since the quality of links depends on changing soil conditions. To address this problem, several path loss models have been proposed to predict the attenuation of an electromagnetic wave in soil. However, this prediction has to be done in-situ by the sensor nodes themselves so that they can avoid wasting energy when transmissions are not possible due to bad soil conditions. In this paper, we propose a link channel optimization for reliable communications in WUSNs based on Sugeno Fuzzy Inference System (FIS). The proposed approach enables a transmitting node to check whether its environment allows it to reliably send data to a receiver. The proposed FIS consists of 4 inputs, one output and 36 inference rules. The inputs give information on the node’s environment, the output gives the probability that data to be sent by a transmitter will be get or not by the receiver. To evaluate the proposed approach, we consider the dataset composed of 140 measurements in dry and moist soil configurations performed at the Cheikh Anta Diop University of Dakar in Senegal. For the validation, we compared our proposal with a recent path loss model called WUSN-PLM according to performance metrics. The results show that our proposal outperforms the WUSN-PLM with higher balanced accuracy (88.21% against 81.061%) and higher Matthews Correlation Coefficient (0.798 against 0.643).