X

Michigan Technological University is making autonomous boats a lot smarter

It's got a tricked-out water scooter at the Detroit Auto Show that will let autonomous surface craft navigate rough seas much more efficiently.

Kyle Hyatt Former news and features editor
Kyle Hyatt (he/him/his) hails originally from the Pacific Northwest, but has long called Los Angeles home. He's had a lifelong obsession with cars and motorcycles (both old and new).
Kyle Hyatt
2 min read
dscn0253
Enlarge Image
dscn0253

Michigan Technical University's water scooter can't run autonomously yet, but once it collects enough data, it should be able to navigate itself through rough seas with ease.

Michigan Technical University

Autonomous cars may be getting most of the ink these days, but autonomous boats are totally a thing too, and the tech behind them is fascinating. The problem with the current state of technology is that it cannot make navigational decisions based on wave conditions and weather in the same way that a human could.

Michigan Technological University is working to change that with a DARPA-funded program designed to navigate rough seas autonomously. The program has been going on for approximately a year and a half and is based at MTU's Marine Autonomous Research Site on Michigan's upper peninsula, and a couple of folks from the program were on hand at the Detroit Auto Show to show off the hard work done by the team there.

The program is currently being developed using a watercraft that is equipped with a nine-axis gyroscope, dual GPS systems, stereoscopic cameras and a whole host of data capturing devices. It's currently able to be piloted as usual by a human onboard or via remote control.
Once sufficient data is collected to accurately predict and recreate a skilled human pilot's decision-making process, the team will focus on making the water scooter fully autonomous. It is essentially a 1/3 scale model of the rigid-hulled inflatable boats that are popular with both the US Coast Guard and the Navy.

Part of what makes programming an autonomous watercraft challenging is the constantly changing surface of the water. A huge number of variables combine to influence wave size and frequency and the MTU team is using the scooter's onboard sensors to help generate accurate wave maps alongside data from a quadcopter drone.

Future incarnations of the tech could use small semi-autonomous short-range sensors that are out in the water and using accelerometers, GPS, and gyros to help the larger vessel plot its course through the waves more effectively. Remember SeaQuest DSV? Think of the "WSKRS" that it used — same idea.

The MTU program is still in its relatively early stages but it's pretty cool, and it goes to show you that sometimes it's worth checking out all the weird stuff in the basement at the Detroit Auto Show.

Roadshow's favorite 2019 Detroit Auto Show debuts

See all photos