Periodic Charging Scheme for Fixed-Priority Real-Time Systems with Renewable Energy

Energy harvesting systems are gaining increasing importance in the embedded systems domain, as they provide an effective solution to bridge the gap between the energy supply and demand. However, the variable nature of the energy supply rate due to the environmental conditions creates serious challenges for embedded real-time systems targeting predictability. This paper presents a proactive and highly predictable framework, called Periodic Charging Scheme (PCS), for fixed-priority real-time systems with renewable energy. The main idea of the algorithm is to plan in advance for periodic charging and discharging of the battery to avoid energy outage, while still meeting the timing constraints. The algorithm is specifically designed to exploit the low-power states of modern processors, to enable effective power state transitions when the battery is recharged on a periodic basis. We also offer online enhancements to opportunistically extend the duration of charging phases and improve the responsiveness of the potential non-real-time workloads without compromising feasibility. Extensive simulations show that the proposed approach outperforms the state-of-the-art algorithms in terms of the number of task sets that meet the timing constraints under specified energy profiles, when a realistic power model with state transition overheads is assumed.