Researchers develop underwater mapping robot
Listen to this article
Underwater environments can be particularly challenging for autonomous robots. Things are constantly moving and changing, and robots need to know where they are without relying on GPS data.
Researchers at Stevens Institute of Technology have created a robot that can successfully navigate a crowded marina underwater. The robot is able to map its environment, track its own position and plan a safe route through a complex environment in real time, simultaneously.
“Underwater mapping in an obstacle-filled environment is a very difficult problem, because you don’t have the same situational awareness as you would with a flying or ground-based robot – and that makes sending a robot underwater is an inherently risky process,” Brendan said. Englot, project manager and acting director of the Stevens Institute for Artificial Intelligence.
The team was able to develop an algorithm that allows the robot to monitor and manage its level of uncertainty about its location and environment, and make real-time decisions based on that uncertainty. The robot uses active SLAM (simultaneous localization and mapping) algorithms.
“Essentially, the robot knows what it doesn’t know, which allows it to make smarter decisions,” Englot said. “By creating a virtual map that takes into account the robot’s own confidence in where it is and what it sees, the robot can quickly, safely, and accurately map a new environment.”
The robot, a custom BlueROV2 robot, operates at a depth of 1 meter and uses sonar signals to detect objects around it. The robot was able to successfully explore a port at the US Merchant Marine Academy in King’s Point, New York.
The robot has many potential applications, including in port repairs, construction and maintenance of offshore wind farms, aquaculture projects and drilling platforms. Going forward, the Englot team plans to strengthen the robotic platform to enable longer duration underwater missions.