Online Direct Policy Search for Thruster Failure Recovery in Autonomous Underwater Vehicles
File(s)
Accepted version
Author(s)
Ahmadzadeh, Seyed Reza
Leonetti, Matteo
Kormushev, Petar
Type
Conference Paper
Abstract
Autonomous underwater vehicles are prone to various factors that may lead a mission to fail and cause unrecoverable damages. Even robust controllers cannot make sure that the robot is able to navigate to a safe location in such situations. In this paper we propose an online learning method for reconfiguring the controller, which tries to recover the robot and survive the mission using the current asset of the system. The proposed method is framed in the reinforcement learning setting, and in particular as a model-based direct policy search approach. Since learning on a damaged vehicle would be impossible owing to time and energy constraints, learning is performed on a model which is identified and kept updated online. We evaluate the applicability of our method with different policy representations and learning algorithms, on the model of the vehicle Girona500.
Date Issued
2013-09
Date Acceptance
2013-09-02
Citation
6th International workshop on Evolutionary and Reinforcement Learning for Autonomous Robot System (ERLARS 2013), in conjunction with the 12th European Conference on Artificial Life (ECAL 2013), 2013
Journal / Book Title
6th International workshop on Evolutionary and Reinforcement Learning for Autonomous Robot System (ERLARS 2013), in conjunction with the 12th European Conference on Artificial Life (ECAL 2013)
Copyright Statement
© 2013 The Authors
Source
ECAL 2013 12th European Conference on Artificial Life
Publication Status
Published
Start Date
2013-09-02
Finish Date
2013-09-06
Coverage Spatial
Taormina, Italy