TY - GEN
T1 - Low Dimensional State Representation Learning with Robotics Priors in Continuous Action Spaces
AU - Botteghi, Nicolò
AU - Alaa, Khaled
AU - Poel, Mannes
AU - Sirmaçek, Beril
AU - Brune, Christoph
AU - Mersha, Abeje
AU - Stramigioli, Stefano
N1 - Publisher Copyright:
© 2021 IEEE.
PY - 2021/12/16
Y1 - 2021/12/16
N2 - Reinforcement learning algorithms have proven to be capable of solving complicated robotics tasks in an end-to-end fashion without any need for hand-crafted features or policies. Especially in the context of robotics, in which the cost of real-world data is usually extremely high, Reinforcement Learning solutions achieving high sample efficiency are needed. In this paper, we propose a framework combining the learning of a low-dimensional state representation, from high-dimensional observations coming from the robot's raw sensory readings, with the learning of the optimal policy, given the learned state representation. We evaluate our framework in the context of mobile robot navigation in the case of continuous state and action spaces. Moreover, we study the problem of transferring what learned in the simulated virtual environment to the real robot without further retraining using real-world data in the presence of visual and depth distractors, such as lighting changes and moving obstacles. A video of our experiments can be found at: https://youtu.be/rUdGPKr2Wuo.
AB - Reinforcement learning algorithms have proven to be capable of solving complicated robotics tasks in an end-to-end fashion without any need for hand-crafted features or policies. Especially in the context of robotics, in which the cost of real-world data is usually extremely high, Reinforcement Learning solutions achieving high sample efficiency are needed. In this paper, we propose a framework combining the learning of a low-dimensional state representation, from high-dimensional observations coming from the robot's raw sensory readings, with the learning of the optimal policy, given the learned state representation. We evaluate our framework in the context of mobile robot navigation in the case of continuous state and action spaces. Moreover, we study the problem of transferring what learned in the simulated virtual environment to the real robot without further retraining using real-world data in the presence of visual and depth distractors, such as lighting changes and moving obstacles. A video of our experiments can be found at: https://youtu.be/rUdGPKr2Wuo.
KW - 22/2 OA procedure
UR - http://www.scopus.com/inward/record.url?scp=85124346914&partnerID=8YFLogxK
U2 - 10.1109/IROS51168.2021.9635936
DO - 10.1109/IROS51168.2021.9635936
M3 - Conference contribution
AN - SCOPUS:85124346914
SN - 978-1-6654-1715-0
T3 - IEEE International Conference on Intelligent Robots and Systems
SP - 190
EP - 197
BT - IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2021
PB - IEEE
T2 - IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2021
Y2 - 27 September 2021 through 1 October 2021
ER -