Imagine you’re a farmer trying to grow a crop under drought conditions. Up-to-the-minute data on soil moisture can help you to decide where and when to irrigate, which directly affects your crop yield and your bottom line. More sensors would mean more data and a better spatial picture of conditions, but the cost of wired soil sensors would be crippling. Wireless sensors that tap into GSM or some sort of mesh network would be better, but each sensor would still need power, and maintenance costs would quickly mount. But what if you could deploy a vast number of cheap RFID-linked sensors in your fields? And what if an autonomous vehicle could be tasked with the job of polling the sensors and reporting the data? That’s one scenario imagined in a recent scholarly paper about a mobile Internet of Things (PDF link).
In the paper, authors [Jennifer Wang], [Erik Schluntz], [Brian Otis], and [Travis Deyle] put a commercially available quadcopter and RC car to the hack. Both platforms were fitted with telemetry radios, GPS, and an off-the-shelf RFID tag reader and antenna. For their sensor array, they selected passive UHF RFID tags coupled to a number of different sensors, including a resistance sensor used to measure soil moisture. A ground-control system was developed that allowed both the quad and the car to maneuver to waypoints under GPS guidance to poll sensors and report back.
Beyond agriculture, the possibilities for an IoT based on cheap sensors and autonomous vehicles to poll them are limitless. The authors rightly point out the challenges of building out a commercial system based on these principles, but by starting with COTS components and striving to keep installed costs to a minimum, we think they’ve done a great proof of concept here.