Increasing Robot Capabilities with AI
The role of artificial intelligence (AI) in technology is to analyze input and provide intelligent, useful, and productive output.
This capability allows technology to take on human tasks in a more efficient and streamlined way, aided by the near limitless power of big data.
The increased need to operate in extreme conditions has accelerated the development of underwater robots. Today, with growing access to AI and machine learning (ML), AUVs and ROVs are used by offshore operators to perform complex subsea tasks and deliver increased service life for oilfield equipment; boosted productivity; enhanced efficiency; improved safety; and reduced operating costs.
Nauticus Robotics, Inc., a developer of ocean robots, autonomy software, and services to the ocean industries, has taken a design approach that combines a robot’s task-performing features in a way that provides more complete solutions for operators with subsea infrastructure.
“We applied concepts from space flight to figure out how to communicate with robotic assets in a remote setting over a latent network,” said Nicolaus Radford, Nauticus Robotics CEO. “We knew that if we cut the umbilical, we’d be limited in the communication techniques available to us and limited in bandwidth. This is why we developed some fairly complex software to help facilitate communications under the sea that increase reliability and drive invariance to latency.”
By recognizing the importance of having a balance between capabilities and controls, Nauticus has designed ROVs and AUVs capable of performing a wide range of manipulative tasks while being controlled by a multi-layered, multi-tool, software platform. This new capability has the potential to transform the way subsea services are provided by creating a platform that is not just for individual robots but unifies all products into a single control architecture to create a “supervised autonomy” for all vehicles across all mission phases, including third-party vehicles.
INSIDE THE TECH
Today, most subsea manipulation tasks require a person to be in direct real-time control of the manipulator. This can be achieved locally or “over the horizon,” but only when certain conditions are met, such as communications being consistently within appropriate speed and latency tolerances.
Nauticus has developed a significantly different approach—one that utilizes an autonomous robot and a human in the loop. The added human supervisory layer of control governs the manipulation process of certain operations to ensure a greater level of safety.
“When directly controlling a robot over a remote connection, there can be some instability and potential for human error,” said Todd Newell, Senior VP of Business Development with Nauticus Robotics. “To solve this problem, we created a machine intelligence framework that can control the robot over a latent and much slower network. In other words, we had to make these robots very smart with a high degree of self-directed and self-sufficient behavior.”
This method of control combines a proprietary acoustic communications method, with a human supervisor who authorizes stepwise autonomous actions through “mouse clicks, not joysticks.” When a robot with advanced control software, powered by AI and ML, is paired with a supervisory feature, the umbilical can be removed, eliminating the need for a large surface support vessel. And, as with any AI/ML software, it gets more capable over time by automatically learning and improving from experience.
The robotic centerpiece of Nauticus’ push into ocean technology is Aquanaut, an all-electric subsea robot with two robust arms that can transform from a long-range AUV to an untethered ROV. In AUV mode, with the arms enclosed within its hydrodynamic hull, the underwater robot can cover over 50 nautical miles in one mission, owing to its onboard lithium-ion battery and thrusters for propulsion.
Along its journey, Aquanaut’s onboard high-precision perception sensors enable it to survey the seabed and collect data. Once the robot reaches its destination, it transforms into ROV mode. The top half of the hull raises and the head swivels into place to expose stereo cameras and powerful 3D sensors.
Additional thrusters emerge to provide better maneuverability. Two eight-axis arms unfold with built-in force sensors and grippers and the robot is ready to perform a range of manipulation tasks. These can include cleaning, inspection of cathodic protection (a common corrosion mitigation method for submerged metal structures), flooded member detection (FMD), and other tasks to assess the condition of seafloor infrastructure.
Aquanaut collects 3D environmental sensory data which is decoded with ML algorithms that help synthesize plans for the robot’s actions within the context of its library of autonomous behaviors. A “human-in-the-loop” directs the autonomous functions via an onshore operator that monitors or manages crucial actions taken by the platform. This provides an additional layer of protection for assets and the surrounding environment.
By using supervised autonomy and the proprietary sea-to-surface network, Aquanaut can operate completely tether-free. For longer projects, Aquanaut can utilize a subsea wireless network to receive instructions and transmit data back to the remote-control center.
Communication is facilitated via an onboard, high-speed acoustic modem. A small unmanned surface vessel relays signals between the robot and communication satellites, enabling the robot to be controlled from anywhere in the world.
Adaptive machine intelligence is at the technological core of the modern robotics landscape, and smarter machines enable a fresh view to ocean-based activities and the ways the human-to-machine interface (HMI) can be changed to provide economic and environmental value.
Starting primarily as a US government contractor, Nauticus Robotics demonstrated how autonomous undersea robots could be utilized to help defend against maritime threats. The company’s leaders then studied how ROVs and AUVs were being used in the O&G industry and knew they had found a niche where they could make a difference. The company embarked on a mission to take the features of these two robotic technologies and combine them into one robotics platform and then remove the need for an umbilical but using machine intelligence and supervised autonomy.
Nauticus Robotics estimates that this technology has the potential to lower the CO2 footprint of offshore vessels by over 90 percent and reduce the need for people offshore by over 95 percent, all at approximately 50 percent of the current cost of traditional methods. Beyond making the oceans safer and offshore energy markets more efficient, Nauticus Robotics is also researching and developing technologies for autonomous shipping, aquaculture, and subsea mining.
For more information, visit: www.nauticusrobotics.com.
To read the full article, which was featured in ON&T September 2022, click here.