Full Program »
Development of Self-aware and Self-Redesign Framework for Wireless Sensor Networks
We propose a self-aware self-redesign framework (SASR), which will embed an existing computing system (CS) with awareness of abnormalities to trigger a self-adaptation process through necessary redesigning to handle the operational challenges. We view wireless sensor network (WSN) deployed in a hostile environment, often subject to gusty winds, as a computing system to be embodied with SASR framework for a seamless transmission at the frequency band of 2.4 GHz. We classified the environment severity into four levels based on the visibility as clear sky, dust haze, dust storm and heavy dust storm, and our framework generates a hybrid-awareness within the sensor nodes by embedding the relationship between the meteorological changes surrounding the sensing area (public awareness) and the corresponding transmission losses (private awareness). Then, by partitioning the transmission band into four channels with varying transmission power and transmission frequencies, and by selecting event and time based channel hopping, the adaptiveness of WSN towards the environment is ensured. We have utilized Contiki’s Cooja simulator, as it suits for low-power and lossy networks, and multi-channel communication, to generate the data transmission with respective transmission path losses for the defined environmental conditions. We have utilized two types of WSN deployment, such as uniform and random and we have attempted the channel hopping mechanism under both uniform and random deployments to test SASR feasibility within WSN. The simulation results showed that uniform deployment, and random deployment with multi-channel hopping based on channel idle level and interference level showed better adaptiveness towards adverse environmental conditions. The simulation results also show a significant improvement in performance metrics in channel hopping under random deployment by showing a drop in packet delivery ratio by 5 to 8%, whereas it is 15 to 30% in random deployment with single channel. It is also further observed that channel hopping is not feasible under uniform deployment for the selected environmental conditions.