The typical controller that human operators use to manipulate a multi-million-dollar undersea robotic arm is a complicated, analog model of the arm itself. Operators direct the movement of each joint, making a costly, high-pressure task such as repairing oil and gas facilities or cleaning up munitions that much harder and more susceptible to mistakes.
Seattle startup BluHaptics has a better way: The familiar video game controller.
The game controller is just one aspect of the software system under development by the six-person University of Washington spinout company, which on Wednesday announced a $1.36 million Series A investment from Seattle-based angel investing groups.
The game controller “is really second nature for anybody who’s grown up playing video games, as almost all kids have,” says BluHaptics CEO Don Pickering. It’s also dramatically cheaper than the industry standard analog master controllers.
It’s the software platform supporting the controller that enables this familiar, inexpensive technology to operate a $5 million robotic arm, and promises to increase efficiency and reduce accidents as well.
While much of the attention in robotics is on factory automation, and increasingly, consumer applications, robots have been used underwater for a couple of decades now, Pickering says.
“The use of tele-operated robots to repair and fix things in the field has really been pioneered in underwater applications—oil and gas and defense—but in the 20 years they’ve been doing that, they haven’t really gone beyond manual control, or analog or manual systems,” he says.
BluHaptics’ software incorporates machine learning, parallel processing, sensor data, and tactile feedback to bring together the best attributes of humans and machines in the operation of robots. Humans can exert their judgment and control capabilities, while the machines provide accuracy and situational awareness—not to mention performing the work itself in dangerous or hard-to-access locations, Pickering says.
The company’s software captures data from sensors, such as high-resolution laser sonar, to create a real-time model of the undersea operating environment. Machine learning helps identify objects. Positional information from the arms themselves contribute to the easier, more intuitive controls: the software automatically translates the pilot’s movements of the game pad to the joints of the arms.
“It’s connecting what a robot sees to what it does,” Pickering says. “There’s a lot of sensors, and there’s a lot of hardware, but we’re really building the first software that closes that loop and allows that information to be used for enhanced and more-precise control, and that is a really big leap from what people are able to do right now.”
The company, founded in 2013 and based on research in the University of Washington’s Applied Physics Laboratory and Electrical Engineering Department, was formed to help the Navy clean up weapons and munitions on the seafloor and lake bottoms, Pickering says. “We realized that there was a much broader application and need for that technology, not just in sub-sea robots, but in terrestrial and space robots,” he says.
BluHaptics aims to have its first commercial product, covering underwater manipulator arms, on the market in June.
“There’s billions of dollars of hardware already out there that could be upgraded with this software,” Pickering says, adding that about a fifth of the company’s code base needs to be tailored to work with a given manufacturer’s hardware—akin to writing printer driver software.
The funding, led by Seattle Angel Fund and joined by Seattle-based Alliance of Angels, will support BluHaptics—which has so far relied mainly on grant funding—as it brings on initial customers and prepares for its first offshore test this month. Pickering says there has already been a strong response, with three early customers in the on-boarding process.
