One-Shot Imitation Learning: A Pose Estimation Perspective
This addresses the problem of enabling robots to learn from minimal demonstrations without additional data, which is incremental as it builds on existing pose estimation methods.
The paper tackles one-shot imitation learning with constraints like a single demonstration and no prior knowledge, formulating it as trajectory transfer and unseen object pose estimation, and evaluates state-of-the-art pose estimators on ten real-world tasks, showing effects of factors like pose estimation error on success rates.
In this paper, we study imitation learning under the challenging setting of: (1) only a single demonstration, (2) no further data collection, and (3) no prior task or object knowledge. We show how, with these constraints, imitation learning can be formulated as a combination of trajectory transfer and unseen object pose estimation. To explore this idea, we provide an in-depth study on how state-of-the-art unseen object pose estimators perform for one-shot imitation learning on ten real-world tasks, and we take a deep dive into the effects that camera calibration, pose estimation error, and spatial generalisation have on task success rates. For videos, please visit https://www.robot-learning.uk/pose-estimation-perspective.