Enabling Reliable Environmental Sensing with LoRa, Energy Harvesting, and Domain Adaptation

Environmental sensing is essential for many applications. Many existing efforts rely on the readings provided by the weather stations maintained by federal, regional, or local government agencies. While the accuracy of the readings provided by those weather stations is high, the ability of such data to reflect the temperature variability experienced by urban populations is generally low. Therefore, recent studies have proposed to deploy new infrastructures with low-power communication and energy harvesting capabilities to provide fine-scale measurements. Recently, there has been an increasing interest in deploying environmental sensing systems with LoRa radios and solar panels. However, there have been very few studies looking into the reliability of solar power in the LoRa-based environmental sensing settings. In this paper, we present an empirical study that investigates how well solar energy powers an environmental sensing platform. Our study shows that solar energy generation forecasting plays an important role in the performance of the sensing platform. To address the challenges, we develop a novel solution that leverages LoRa, energy harvesting, and domain adaptation to enable reliable environmental sensing. Experimental results show that our solution outperforms the baselines and effectively supports end devices to perform environmental sensing operations without interruptions.