The work done in the article was published in the journal robotics and autonomous systems.

About the article

In this work we develop a deep RL framework for adaptive control of AUVs based on the well known deep deterministic policy gradient algorithm (DDPG). The proposal was a goal-oriented deep RL architecture, which takes the available raw sensory information as input and gives as output the continuous control actions which are directly the low-level commands for the AUV’s thrusters.

The research was perform at the Ocean Systems Laboratory of the Heriot-Watt University, using Nessie, the autonomous underwater vehicle developed by students and researchers of the university.