The vegetable garden at Helen Greiner’s oceanfront home in Beverly is surrounded by a tall steel fence. It keeps the squirrels and chipmunks out, and keeps the robots in.
The robots don’t seem to mind. The pair of bright green, disk-shaped machines trundle randomly among the tomato seedlings and cabbage sprouts, churning the soil with wheels that destroy unwelcome weeds, so Greiner doesn’t have to.
“It’s one of those hated chores, so it makes sense to have a robot do it,” said Greiner, chief executive of Tertill, the Billerica company that makes the automated gardeners.
In Greiner’s world, robots always get the dirty jobs. After graduating from MIT in 1990, she cofounded iRobot, the Bedford-based company that began life as a maker of battle-hardened military droids. In 2002, iRobot introduced Roomba, a robot vacuum cleaner that became a global hit, with over 30 million sold. To this day, neither iRobot nor anybody else has created a home robot anywhere near as popular. It’s as if consumers won’t trust robots to do anything but clean their floors.
But Greiner thinks the world is ready for a $349 machine that promises to wipe out weeds as easily as a Roomba sucks up dust bunnies.
It’s a gamble, but like all serial entrepreneurs, Greiner has a certain tolerance for risk. She revels in the nor’easters that pummel the nearby shoreline. On calm days she climbs into a kayak and paddles for miles across Salem Sound. She’s visited every island in sight — Eagle Island, Bakers Island, Little Misery.
Greiner had a sure thing with iRobot, but walked away in 2008. She then launched and sold the aerial drone developer CyPhy Works, leaving that company in 2018 for a two-year stint as a robotics advisor for the Army.
Then in 2020, Greiner spotted a posting on the fund-raising site Kickstarter by an old friend from iRobot — Joe Jones, the engineer who’d designed the original Roomba. Jones had left iRobot to join Harvest Automation, a builder of robots for commercial agriculture. But around 2015 a friend suggested that Jones return to his roots. Just as he’d once built a robot to clean up living rooms, why not a machine to tidy up gardens?
Jones had built his first Tertill prototype by 2016, and was pleasantly surprised by how well it worked. After multiple upgrades, he ended up with a sort of garden Roomba, encased in bright green plastic with a square solar panel on top. In 2017, Jones launched the Kickstarter campaign to manufacture the machine.
Greiner took note and decided to purchase one. “I didn’t expect it to work,” she said. “But I put it in my garden and it did a phenomenal job.” In 2020, Greiner signed on as chief executive of the young company, with Jones as chief technology officer.
Like the original Roomba, the Tertill lacks a powerful processor or a costly guidance system. “I think that the smartest robots are the ones that have enough intelligence to do the job at the lowest possible price,” said Jones. In his iRobot days, he found that if the Roomba bounced off living-room obstacles and kept on moving, it eventually cleaned every inch of the floor.
A Tertill works in roughly the same way. Its sensors can identify and avoid relatively large objects — fences, or 4-inch-high tomato plants — and steer around them. It comes with little metal “fences” that can be shoved into the earth to shield newly planted seedlings. These fences are important, because any plant that the Tertill is small enough to roll over will be doomed. The robot’s capacitive sensor detects the moisture in plant leaves and stems. Then it fires up a belly-mounted mini-weed whacker that rips it apart.
Yet the weed whacker is only a last resort. The wheels do most of the work; by thrashing the soil, they tear up newly sprouted weeds before they can take root. “That was a big surprise to us,” said Jones. “The robot itself had discovered a whole new way to destroy the weeds.”
Each Tertill can tend up to 200 square feet of garden space. It’s weatherproof, so it can stay outdoors throughout the growing season. A Tertill runs for one to two hours per day, for maybe five minutes at a time. Then the robot pauses, to let its solar power cell recharge the battery.
There are plenty of companies developing robot gear for large-scale agribusiness, but Tertill has the backyard mostly to itself. Perhaps its prime competitor is FarmBot, a California company that makes a robotic system that’s mounted above a vegetable patch and rides on rails. Modeled after computer-controlled machine tools for manufacturing, the FarmBot can choose the correct tool for a variety of tasks — planting seeds, watering, or weeding. It’s a far more sophisticated machine than Tertill, but at a far higher price — between $3,500 and $5,000.
Greiner and Jones aren’t trying to build anything nearly so ambitious, but they are toying with some possible upgrades. Pest control, for instance. Jones has built a prototype that uses an infrared sensor to detect the body heat of hungry mammals. Then the robot rolls toward the critter, scaring it away.
“I have videos of bunnies running away,” said Jones.
Greiner won’t say how many Tertills have been purchased so far. But then, how many consumers will line up to buy a weed-killing, squirrel-chasing robot?
Jonathan Collins, a smart-home analyst for ABI Research, thinks the Tertill has a fighting chance. After all, nobody thought about buying robot vacuum cleaners, until the Roomba proved its worth. “I think it’s about price point, reliability, and customer education,” said Collins. “These things spread by word of mouth and experience.”
In the meantime, Tertill has rolled out a new service that tests a customer’s garden soil, then ships them a customized blend of organic fertilizer to produce optimum results. At prices ranging from $59 to $99, the Tertill Garden Program is intended for gardens ranging from 32 to 120 square feet.
It’s a clever idea, but one that’s surprisingly free of electronics. No matter, said Jones. “The idea for the company is to make gardening easier and more attractive,” he said. Sometimes that means turning a robot loose in the backyard, but sometimes it doesn’t.