A Fuzzy Framework for Self-Aware Wireless Sensor Networks

Abstract

We have developed a fuzzy logic framework to simulate a self-awareness within wireless sensor network (WSN) that envisages data transmission losses due to abnormalities in the operating environment. Fuzzy logic is selected to model the awareness, as it can be coded using less data, and occupies less memory, thus, suited to resource-constraint sensor nodes. We view WSN as a multi-layered infrastructure, where the top-layer comprises of a central server managing the intermediate-layers’ gateways, which are aggregating the raw sensed data from the bottom-layer sensors. The self-awareness is simulated using Mamdani Fuzzy inference system (FIS) that utilizes a custom-defined membership functions of four parameters, such as temperature, humidity, wind speed, and battery residual-level to guide WSN through its operations. The input and output parameters are defined with three linguistic states (Low, Medium and High) and a defined set of rules are applied to the selected FIS to assess the impact level. The derived impact is then mapped into transmission efficiency using packet error rate-environmental uncertainties relationship and the gateway with the awareness of selected uncertainties decides on executing the data transmissions. A multivariate dataset is generated by combining real time meteorological data with randomly generated battery voltage and the experimental results on the dataset revealed that the proposed FIS framework demonstrated 33% energy savings by non-executing unworthy communications against normal transmissions for WSN deployed outdoor.