Position: drone camera communication meets robotic soil sensing

Abstract

This paper positions the idea of enhancing the sensing capabilitiesof unmanned aerial vehicles (UAV or Drones) using LED-Cameracommunication, camera perception, and adaptive sampling algo-rithms. In this paper, considering soil moisture measurement overa specific geographic area as the application, we propose a systemthat consists of drone assisted mobile ground robots(MGRs) thatperform collaborative sensing and communicate with the droneusing visible light. The key idea is to combine aerial images andMGR sensed moisture values to generate a dynamically adaptablemoisture map. Camera communication eliminates the need to estab-lish continuous radio communication in remote locations and offersvisual association that enables the drone to localize the MGRs. Thissystem pushes the state-of-the-art in sensing by accurately sensinga region with high spatio-temporal resolution while optimizing thetime and energy required to create a moisture map of a geographi-cal area. As the sensing resolution is dynamic, we hypothesize thatthe proposed system can efficiently predict, detect, and map soilmoisture hotspots.