Dynamic harvesting- and energy-aware real-time task scheduling

Energy harvesting, along with effective storage of the energy, is a very common approach to attain sustainable computing in today's embedded systems. Employing a hybrid energy storage system (HESS), which constitutes of two or more types of energy storage systems (ESSs), helps to compensate for the weaknesses of one ESS type using the strengths of another type. The capacity of an ESS, and thus that of a HESS, can be modeled by dividing it into Instantly Available Charge (IAC) and Instantly Unavailable Charge (IUC) parts; the existing charge in an ESS always flows from the part with higher voltage to the other one. The main idea of this study is to intelligently control the flows in the HESS to maximizing either the IAC or the IUC charge. We propose the HLPF real-time task scheduling algorithm to do so through deciding to execute the tasks in the ascending or descending order of their power requirements. Extensive simulations show impressive lifetime improvements of up to 20 % in comparison to the classical real-time task scheduling algorithms. (C) 2020 Elsevier Inc. All rights reserved.