When underwater researchers use autonomous underwater vehicles (AUV), they find that the AUVs really aren’t that autonomous. Researchers have to program all of the movements that the AUV makes, making for a long process of individual instructions. However, researchers at the Massachusetts Institute of Technology (MIT) have now designed a new algorithm that allows underwater robots to program themselves essentially, and make travel decisions on its own based on basic parameters.
High-Level Decision Making Prowess of Underwater Robots
A principal in the development of the program, MIT Professor Brian Williams, states that the underwater robot with the new programming can execute a mission and make adaptations without any help from a human. The recent report was published in the MIT Technology Review. Given basic instructions from an engineer, the underwater robots decide how best to accomplish its given task. For instance, all the engineer needs to do is provide a list of goals, time parameters and any physical directions, such as staying at a certain level above the ocean floor, and the robot will do the rest by itself. In the event something unforeseen occurs, the underwater robot has the ability to make the decision to drop a particular task, or to reprogram its moves to accommodate the changes, all without human input.
Freeing Researchers for More Important Decisions
Allowing underwater robots a high level of decision-making frees up researchers to concentrate on overall research strategy and reduce the number of people needed for underwater research teams.
Current research generally includes only one underwater robot because of the programming requirements and the number of people required to watch over the machines. However, with AUVs able to make their own decisions, a group of them could be used, each one going about its mission and travelling clear of each other.
Successful Testing Completed in March
The MIT team conducted a series of tests off the coast of Australia in March. Working together with Woods Hole Oceanographic Institution in Massachusetts and the Australian Center for Field Robotics, tests took place over a period of three weeks. Testing showed that the new algorithm performed exactly as hoped, with the underwater robot making high-level decisions for its mission. It was also shown that other similarly programmed robots basically cooperated with each other without the need for human instructions.
Professor Williams stated that in the future, it might be unnecessary to have humans along for deep-sea exploration. With a system of these types of underwater robots working together, they can zigzag across the ocean floor like going through an obstacle course, and travel along completing their missions unaided. Future AUVs will have the ability to decide if a task is too dangerous and pass it up or postpone, and even be able to make its own repairs and then continue on about its business.
