Sheep graze the University of Nevada, Reno's main station field lab. They trod along the same worn dirt paths to reach their water stations. Around the troughs, the soil is dry, and the grass is sparse.
Andrew Hess, a sheep geneticist at UNR, said those sheep grazing patterns are causing low-quality pasture around the troughs.
"If you keep the tanks in a fixed location, that results in a path that the animals typically take to get there, which means that the pasture can get worn in those places," Hess said.
Hess hopes a new tool — called RoboHydra — can help change that. It's a mobile robot with a closed watering tank and built-in health monitoring. Engineers are in the process of developing the prototype.
Water management for arid climates is just one of RoboHydra's goals.
"Water is a critical resource," Hess said, "Especially in Nevada, where we experience harsh droughts and things like that, water is always a key necessity."
The robot will roam the field, providing water to the grazing sheep. The idea is that the roaming robot will vary sheep grazing patterns. The robot will have safety features to ensure this is done with the safety of both the sheep and the robot in mind.
Robohydra will also help track their health. Parikshit Maini, a roboticist at UNR, said the robot will use cameras and thermal imaging to spot issues early.
"So by doing this health monitoring, we can potentially identify health issues that might say a sheep is losing weight, for example," Maini said, "that gives a cue to the farmer to maybe take a look at this sheep to see if there's any issues that you can catch before it becomes a bigger problem and drive up [veterinary] costs."
The system will also collect data to help researchers understand which sheep are better suited for dry environments. Images of the animals will be fed into an AI model to identify patterns associated with water use and heat resilience. This will be combined with the genetic data Hess acquires.
Maini said the robot will help producers decide which sheep to breed.
"I was talking to one of my colleagues down the corridor, and he was like, 'Are you making sheep Tinder?'" Maini laughed.
Ankita Shukla, leading the AI side of the project, hopes it leads to more.
"I think if I look down maybe five, six years down the line, the aim would be to make the system accessible," Shukla said, "With the current setup … we just started scratching the surface with this idea."
Although agricultural robotics is not new, this project is the first of its kind in the sheep industry. The end goal, after four years, is to give producers better tools to make decisions in an unpredictable climate.
