Suelo: human-assisted sensing for exploratory soil monitoring studies

  • Authors:
  • Nithya Ramanathan;Thomas Schoellhammer;Eddie Kohler;Kamin Whitehouse;Thomas Harmon;Deborah Estrin

  • Affiliations:
  • Center for Embedded Networked Sensing;Cisco / Ironport Systems;UC Los Angeles / Meraki;University of Virginia;UC Merced;Center for Embedded Networked Sensing

  • Venue:
  • Proceedings of the 7th ACM Conference on Embedded Networked Sensor Systems
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Soil contains vast ecosystems that play a key role in the Earth's water and nutrient cycles, but scientists cannot currently collect the high-resolution data required to fully understand them. In this paper, we present Suelo, an embedded networked sensing system designed for soil monitoring. An important challenge for Suelo is that many soil sensors are inherently fragile and often produce invalid or uncalibrated data. Therefore Suelo is an assisted sensing system: it actively requests the help of a human when necessary to validate, calibrate, repair, or replace sensors. This approach allows us to use available sensors without sacrificing data integrity, while minimizing the human resources required. We tested our system in multiple real soil monitoring deployments and demonstrate that, using human assistance, Suelo produced 91% fewer false negatives and false positives than common fault detection solutions on these datasets.