Low-Cost Reliable Blackout Sustainability of Wireless Sensor Networks with Energy Harvesting Systems

Konferenz: European Wireless 2011 - Sustainable Wireless Technologies
27.04.2011 - 29.04.2011 in Vienna, Austria

Tagungsband: European Wireless 2011

Seiten: 8Sprache: EnglischTyp: PDF

Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt

Autoren:
Glatz, Philipp M.; Hörmann, Leander B.; Steger, Christian; Weiss, Reinhold (Institute for Technical Informatics, Graz University of Technology, Graz, Austria)

Inhalt:
Wireless sensor network (WSN) motes are typically small and networked embedded systems. They are of limited computational power and suffer severe resource constraints - especially energy is scarce. With the advent of energy harvesting system (EHS) technology new opportunities open up for long-lived WSNs. Several prototypes of EHS-enhanced WSN platforms have recently been designed in a low-power and energy-efficient way. However, while the field starts to mature there are still a few lessons to be learned for building low-cost and robust systems. While the term low-cost may easily be defined by means of calculating what hardware and maintenance cost per end-user performance for a given business model, it is more complicated to define what is meant by the term robust. For perpetual systems one has to find a formal way of validating whether a system implementation - including all low power hardware and networking approaches - can sustain a given quality of service despite harsh environmental conditions or faulty operation. For the presented case study with the RiverMote hardware platform, we model its black-out sustainability (BOS). We can validate that with its robust low-power design and BOS of more than three weeks can be achieved. Perpetual operation with error detection, fall-back mechanisms, error reporting and self-healing of the system is possible. The system relies on more easy-to-validate system-level mechanisms than other approaches that concentrate on energy harvesting device pattern prediction and therefore have to assume robust forecasting of environmental conditions.