We introduce a deep reinforcement learning framework that enables the generation of trained policy networks (agents) that incorporate local knowledge of current fields to provide energy-minimizing guidance for unmanned underwater vehicles (UUVs). Agents trained using this framework are capable of generating heading, depth, and speed commands that yield nearly-optimal, minimum-energy transits through spatially varying current fields. Numerical simulations in two and three spatial dimensions demonstrate that the energy cost of transits executed under the guidance of a trained agent are better than dead-reckoning and approach known optimal performance in a set of test scenarios.
Reinforcement Learning for Improved Guidance and Power Management of Unmanned Underwater Vehicles
2023-09-25
585115 byte
Conference paper
Electronic Resource
English
Feature-Related Acoustic Guidance of Unmanned Underwater Vehicles
British Library Conference Proceedings | 2002
|Modeling, Identification, Control and Guidance of Unmanned Underwater Vehicles
British Library Conference Proceedings | 2001
|A review of guidance laws applicable to unmanned underwater vehicles
Tema Archive | 2003
|Motion control of unmanned underwater vehicles via deep imitation reinforcement learning algorithm
Wiley | 2020
|Motion control of unmanned underwater vehicles via deep imitation reinforcement learning algorithm
IET | 2020
|