May 21, 2021
In many domains, the goal for autonomous system development is to match human performance. In others, autonomous systems extend human reach, providing capabilities unachievable by humans or human-occupied platforms. The ocean provides a powerful example of the latter domain. Marine robots are transforming how we explore, work, and fight in the marine environment. The history of marine robots thus provides useful insights to the challenges we face as robotic systems shift from replicating humans to carrying out activities we never imagined. A primary lesson from marine robotics is that reliability requires intelligence. Although the robotics literature is full of work focused on negotiating the challenges of a complex environment, for a robot that operates beyond human assistance and without ready communications, the challenge is internal as much as external. This problem is particularly acute for marine robotics given the high cost of ocean testing and the dearth of appropriate test facilities. A second important lesson is the value of numbers. Multi-platform approaches to observing and working in the ocean are becoming commonplace with the advent of robotic systems that cost a fraction of human-occupied platforms like ships or submarines. While the mathematical advantages are easily calculated, the timelines and costs of assuring system performance are multiplied many-fold as the complexity of the built system increases and the environment effectively becomes part of the ‘natural’ system in many implementations. Validation and verification of our systems are now the bottleneck, absorbing an ever-larger fraction of our development effort. Often, key insights emerge during final tests when the system redesign comes at great cost. Cumulatively, this suggests that a broad engagement in the science, technology, and engineering of assurance of autonomous systems, and of autonomous systems of systems, will reap great advantages. These efforts will yield both better systems and a faster development cycle. Although the technical challenges are severe, the benefits should reward the efforts many times over.
James G. Bellingham is a pioneer in the development of autonomous marine robots. He has led and participated in research expeditions around the world from the Arctic to the Antarctic. Jim is the newly appointed Executive Director of the Institute for Assured Autonomy at Johns Hopkins University. Since 2014, he previously served as the founding Director of the Center of Marine Robotics at the Woods Hole Oceanographic Institution (WHOI). He founded the Autonomous Underwater Vehicles Laboratory at MIT (1988), and co-founded Bluefin Robotics (1997). He was Director of Engineering and Chief Technologist at the Monterey Bay Aquarium Research Institute (MBARI) before coming to Woods Hole.
Dr. Bellingham has a particular interest in marine robotics entrepreneurship, working with startups in a range of capacities and hosting community building events such as the Marine Robotics Entrepreneurs Forum. Jim serves on numerous advisory committees and boards. He Chaired the Naval Research Advisory Committee for five years and serves on the Naval Studies Board and has been on numerous National Academies studies. Jim’s awards include the Lockheed Martin Award for Ocean Science and Engineering, the MIT Fourteenth Robert Bruce Wallace lecturer, and the Department of the Navy Superior Public Service Award. Jim was recently elected as a member of the National Academy of Engineering.
Jim received an S.B. and Ph.D. in physics from the Massachusetts Institute of Technology.