This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).
Many sensor network deployments are in harsh terrain that lack infrastructure but also need to use solar energy harvesting for self-sustained operation. The design of harvesting-aware solar sensor networks raises numerous systems and networking challenges. From a systems perspective, a design should be able to operate under a wide dynamic range of energy harvesting scenarios, from plentiful sunlight to very limited light (e.g., thickly forested areas). From a network perspective, a protocol stack should be able to adapt to spatio-temporal variability in harvesting rates due to foliage and day-to-day vagaries.
In this project, we will design systems and network support for harvesting-aware solar sensor networks. Our systems contributions include novel sensor platforms that use thin-film batteries for harvesting under low-light conditions, predictive duty-cycling under highly variable harvesting conditions, and lightweight checkpointing and restart to handle intermittent loss of power. Our networking contributions include a link layer that is optimized for duty-cycle variability, a network layer that uses harvesting-aware path metrics, and a transport layer can offer high throughput under diverse harvesting conditions.
Our educational plans consist of K-12 teaching through a summer high school outreach program, and inter-disciplinary REU programs in collaboration with Harvard Forest for deploying solar-powered sensor networks for ecological monitoring.
