Optimization of Environmentally Powered Wireless Sensor Networks for Efficient Energy Harvesting

Abstract

Focusing on environmentally powered Wireless Sensor Networks (WSNs), this thesis studies optimized operation of individual sensor node in terms of average duty cycle. In particular, the focus lies in gaining high average duty cycle with high stability. To achieve this objective, an energy neu- tral approach based efficient power management system is introduced and investigated in different working conditions. WSNs deployed in ad hoc manner comprise of numerous sensing nodes organised in a network for the sake of checking and balancing the environmental factors. Each node has sensing, computation, communication and locomotion capabilities but operates with limited battery life. Energy harvesting is a way of powering these WSNs by harvesting energy from the environment. Using harvesting energy as source, certain considerations are different from that battery operated networks. Nondeterministic energy availability with respect to time is the reason behind these differences which put a limit on the maximum rate at which energy can be used. Thus, power management is of prime importance in self-powered networks. The thesis begins with development of efficient solar forecasting algorithm for accurate estimation of energy availability. Reliable knowledge of solar radiation is essential for in- formed design, deployment planning and optimal management of energy in rechargeable sensor networks. In the proposed work, an optimized Pro-Energy algorithm is developed using level and trend factors in time series analysis for future solar irradiance estimation. The performance of proposed algorithm has been compared with EWMA, WCMA, and Pro-Energy on the basis of prediction error. The problem of solar irradiance forecasting has been further addressed by machine learning methodologies over historical data set. In proposed work, forecasts have been done using FoBa, leapForward, spikeslab, Cubist and bagEarthGCV models. To achieve more precise and consistent forecast, four Sta- tistical Ensemble (SE) approaches have been presented.