Self-Supervised Domain Adaptation for Visual Navigation with Global Map Consistency
Open Access
- 1 January 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1868-1877
Abstract
We propose a light-weight, self-supervised adaptation for a visual navigation agent to generalize to unseen environment. Given an embodied agent trained in a noiseless environment, our objective is to transfer the agent to a noisy environment where actuation and odometry sensor noise is present. Our method encourages the agent to maximize the consistency between the global maps generated at different time steps in a round-trip trajectory. The proposed task is completely self-supervised, not requiring any supervision from ground-truth pose data or explicit noise model. In addition, optimization of the task objective is extremely light-weight, as training terminates within a few minutes on a commodity GPU. Our experiments show that the proposed task helps the agent to successfully transfer to new, noisy environments. The transferred agent exhibits improved localization and mapping accuracy, further leading to enhanced performance in downstream visual navigation tasks. Moreover, we demonstrate test-time adaptation with our self-supervised task to show its potential applicability in real-world deployment.Keywords
All Related Versions
Funding Information
- National Research Foundation of Korea
This publication has 24 references indexed in Scilit:
- Globally-Optimal Inlier Set Maximisation for Camera Pose and Correspondence EstimationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2018
- Matterport3D: Learning from RGB-D Data in Indoor Environments2017
- Cognitive Mapping and Planning for Visual Navigation2017
- Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception AgeIEEE Transactions on Robotics, 2016
- ORB-SLAM: A Versatile and Accurate Monocular SLAM SystemIEEE Transactions on Robotics, 2015
- Application of Dijkstra algorithm in robot path-planning2011
- Sensor and actuator modeling of a realistic wheeled mobile robot simulator2008
- Simultaneous localization and mapping: part IIEEE Robotics & Automation Magazine, 2006
- A frontier-based approach for autonomous exploration2002
- Categorization of parameters in the dynamic robot modelIEEE Transactions on Robotics and Automation, 1989