Georgia Institute of Technology
UNIVERSAL NAVIGATION: LEARNING A GENERAL NAVIGATION POLICY FOR HETEROGENEOUS ROBOTS
Abstract
dc:description.abstractTarget-driven visual navigation is a challenging problem that requires a robot to find the goal using only visual inputs. Many researchers have demonstrated promising results using deep reinforcement learning (deep RL) on various robotic platforms, but typical end-to-end learning is known for its poor extrapolation capability to new scenarios. Therefore, learning a navigation policy for a new robot with a new sensor configuration or a new target still remains a challenging problem, which could be defined as a ’Universal Navigation’ problem. The objective of the proposed research is to find a universal policy for the agent to quickly adapt to new sensor configurations or target objects, and successfully navigate in unseen situations. In this project, we design a policy architecture with latent features between perception and inference networks and quickly adapt the perception network via meta-learning while freezing the inference network. Our experiments show that our algorithm adapts the learned navigation policy with only three shots for unseen situations with different sensor configurations or different target colors. We also analyze the proposed algorithm by investigating various hyperparameters. A paper based on this work was accepted to International Conference on Robotics and Automation(ICRA) 2021.
Degree
thesis:*- Level thesis:degree_level
- Masters
- Department dc:contributor.department
- Electrical and Computer Engineering
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2021
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Luo, Qian
- Advisors dc:contributor.advisor
-
- Ha, Sehoon
- Bloch, Matthieu R.
- Committee member dc:contributor.committeemember
-
- Gombolay, Matthew
Subjects
dc:subject × 1Identifiers
dc:identifier.*- Handle dc:identifier.uri
- http://hdl.handle.net/1853/65133