Vintners in California’s Central and Napa valleys are facing two shortages: of water and of workers. What’s more, the state’s recent drought has caused a financial strain for growers thanks to the cost of irrigating their thousands of hectares of vines. And stepped-up enforcement of immigration laws, competition from less strenuous higher-paying jobs, and an aging workforce have led to a labor shortage, according to Wine Spectator magazine.
When researchers from the University of California met with grape growers to discuss how technology could help them, the growers asked if their irrigation systems could be made more efficient and with less human intervention.
“I’m a roboticist, so whenever there’s a problem, I study whether a robot could solve it,” says IEEE Senior Member Stefano Carpin, a professor of electrical engineering and computer science at UC Merced.
Carpin leads a team of researchers from UC Berkeley, UC Davis, and UC Merced who are building the Robot-Assisted Precision Irrigation Delivery (RAPID) system. The system involves inexpensive, adjustable, plastic water emitters attached to the holes in drip irrigation lines. Each emitter would help regulate the amount of water discharged. And because there aren’t enough workers to adjust the hundreds of emitters each vineyard is expected to need, the group is designing rugged, battery-operated robots to move around and handle the job.
The project, launched in 2016 and funded by a nearly US $1 million grant from the U.S. Department of Agriculture, is part of the National Science Foundation–led National Robotics Initiative.
“This robotic system would make fine-grain adjustments to deliver water on a per-plant level,” Carpin says. “We would reduce water use while preserving the quality of the grapes.”
Growing grapes is a tricky business. Their quality depends in part on how much water is absorbed by the vines’ roots. Current drip-irrigation systems consist of a rubber pipe pierced with holes that is laid down along a row of vines. Ideally, vineyards should apply so-called localized stress irrigation, which customizes the amount of water delivered to each vine. But most drip-irrigation systems can’t deliver water with precision. The drip pipes have emitters spaced from 300 millimeters to 900 mm apart—which can’t be adjusted.
Water dripping through the holes wets the fields in uniform blocks, often of around 64 hectares. “There’s a lot of variability in these blocks in terms of soil moisture, microclimate, and weather conditions, so delivering the same amount of water in a fixed flow to the whole block is not optimal,” Carpin says. “Some grapes end up with too much, some with too little.”
With RAPID, each watering pipe is, in effect, turned into a precision irrigation system.
The team is exploring a plastic emitter with a cap that can be turned by a robot to adjust the water flow. Each emitter would cost about 20 cents.
“It’s a simple system,” Carpin says. “Turn the cap clockwise to increase the flow and counterclockwise to decrease it.” He says he can’t estimate how much water the system might save until it has been tested.
To help determine when an adjustment is needed, RAPID will rely on data from field-monitoring systems such as drone and satellite imagery, weather satellites, and services including the California Irrigation Management Information System. CIMIS is a network of more than 145 automated weather stations to help irrigators manage their water.
Carpin is trying out the four-wheel Husky ground vehicle from Clearpath Robotics—basically an open-topped box that carries the elements of the robot. He expects his finished vehicle will be slightly larger than the Husky to accommodate batteries and an arm with a grasping hand. The robot also will have a GPS receiver to map its route and an RFID reader on board to direct the machine to the emitter, he says. The robot will position itself and with its grasping hand turn the cap to adjust water flow, he says, guided by readings from a monitoring system plugged into its flash drive.
By next year, Carpin says, he will have a prototype of his system, and an irrigation pipe pierced with holes will be retrofitted with the adjustable emitters. He expects the system to be tested on a farm, which has already been selected, by the summer of 2020.
Carpin estimates that initially each system will cost in the tens of thousands of dollars, depending on the size of the farm.
You can read about an earlier version of RAPID in “DATE: A Handheld Co-Robotic Device for Automated Tuning of Emitters to Enable Precision Irrigation,” available in the IEEE Xplore Digital Library. Carpin and his co-authors plan to present papers about RAPID at the International Conference on Robotics and Automation, taking place in Brisbane, Australia, from 21 to 25 May.
This article is part of our April 2018 special issue on agtech.