Light Insight — Emulation of Radiation Traces for Analysis and Evaluation of Solar-Harvesting Algorithms
International Workshop on Energy Harvesting and Energy-Neutral Sensing Systems, SenSys 2017: 8-13 (2017-11-05)
Contribution to Conference
The restricted energy budget of energy-aware and -predictive sensor nodes demands algorithms for adaptive duty-cycling and task scheduling. However, their comparison and development is hindered by the lack of reproducibility of environmental conditions, such as solar radiation. To debug their behavior efficiently, it is also key to replay conditions in which the algorithms react unexpectedly in the laboratory. To investigate real-world issues such as dust or pollen, we enable replaying recorded lighting conditions by building an affordable light box. Our self-developed control circuit and high-power LEDs allow us to repeatedly replay environmental illumination data through previously collected and also artificial current and voltage traces. By using the on-board ADC of the harvester, we are able to replay the conditions exactly as seen by the sensor node. We achieve replaying accuracy below the limits of the sensor node’s sensing circuitry, which enables us to directly compare the behavior of nodes running different energy-aware and -predictive algorithms.