DRINKING IN THE DATA
The UC Merced team has partnered with a major commercial grape grower with vineyards near Merced and Madera. During the past two summers, team members collected data using soil moisture probes and drones equipped with thermal and multispectral sensors to measure foliage color and how many leaves are on the vines — both indicators of water needs.
Graduate student Thayer is now using all that data to create algorithms to optimize and instruct where the robots or field workers should make irrigation adjustments.
The robots would be equipped with customized grippers to manipulate inexpensive plastic emitters. The workers would be equipped with wireless, cloud-connected hand-held devices feeding them the same information.
The RAPID team’s work already has earned major recognition. In August, Thayer was the lead author of a paper titled “Multi-Robot Routing Algorithms for Robots Operating in Vineyards” that won the Best Conference Paper Award at the 2018 IEEE International Conference on Automation Science and Engineering (CASE) in Munich, Germany.
“People are inclined to be interested in things that involve wine and robots,” he joked about the honor. More seriously, he added, “It’s an extremely difficult problem to solve in an ordinary case — and for a case like ours, where we have a huge vineyard with 100,000 data points to process, it just makes the problem that much more difficult. Being able to create a solution for that is a major accomplishment.”
People are inclined to be interested in things that involve wine and robots.
UC Merced Graduate Student Thomas Thayer
Among the challenges, Thayer said, is figuring out how to route multiple robots, with finite battery life, through a large vineyard where rows are long and movement is limited.
The team is heading back to the fields this year to apply and test the algorithms, albeit without the irrigation system in place yet. Meanwhile, UC Berkeley partners are designing the gripper that will be mounted on the robot arm and used to adjust irrigation emitters.
Professor Goldberg of UC Berkeley, director of the People and Robots Initiative at the Center for Information Technology Research in the Interest of Society (CITRIS) and the Banatao Institute, is looking forward to his next step: moving out of digital simulation and into the dirt.
His team recently created a small irrigated garden planted with drought-sensitive grass, so the gripper can be tested in real life.
“If we can do this convincingly on a small scale, we can approach companies that are developing agricultural robotics and with that evidence be able to get them on board to partner with us to then to start testing it in real fields,” Goldberg said.