Reinforcement Learning Helps Underwater Robots To Locate And Track Objects

A research team led by the Institute of Marine Sciences (ICM-CSIC) in Barcelona, featuring professor Mario Martin from the UPC’s Department of Computer Science, has demonstrated for the first time that deep reinforcement learning enables underwater autonomous vehicles and robots to accurately locate and track objects and marine animals. This technology makes it possible to study small-scale phenomena, such as CO₂ capture by marine organisms, which helps to regulate climate change. The details of this work are reported in a paper published in Science Robotics.