Visual Imitation Learning of Non-Prehensile Manipulation Tasks with Dynamics-Supervised Models
Authors:
- Abdullah MustafaAMAbdullah Mustafa
- Ryo HanaiRHRyo Hanai
- Ixchel RamirezIRIxchel Ramirez
- Floris ErichFEFloris Erich
- Ryoichi NakajoRNRyoichi Nakajo
- Yukiyasu DomaeYDYukiyasu Domae
- Tetsuya OgataTOTetsuya Ogata
Preprint
Abstract
Unlike quasi-static robotic manipulation tasks like pick-and-place, dynamic tasks such as non-prehensile manipulation pose greater challenges, especially for vision-based control. Successful control requires the extraction of features relevant to the target task. In visual imitation learning settings, these features can be learnt by backpropagating the policy loss through the vision backbone. Yet, this approach tends to learn task-specific features with limited generalizability. Alternatively, learning world models can realize more generalizable vision backbones. Utilizing the learnt features, task-specific policies are subsequently trained. Commonly, these models are trained solely to predict the next RGB state from the current state and action taken. But only-RGB prediction might not fully-capture the task-relevant dynamics. In this work, we hypothesize that direct supervision of target dynamic states (Dynamics Mapping) can learn better dynamics-informed world models. Beside the next RGB reconstruction, the world model is also trained to directly predict position, velocity, and acceleration of environment rigid bodies. To verify our hypothesis, we designed a non-prehensile 2D environment tailored to two tasks: "Balance-Reaching" and "Bin-Dropping". When trained on the first task, dynamics mapping enhanced the task performance under different training configurations (Decoupled, Joint, End-to-End) and policy architectures (Feedforward, Recurrent). Notably, its most significant impact was for world model pretraining boosting the success rate from 21% to 85%. Although frozen dynamics-informed world models could generalize well to a task with in-domain dynamics, but poorly to a one with out-of-domain dynamics.Keywords
All Related Versions
This publication has 4 references indexed in Scilit:
- Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots2024
- Keep It Upright: Model Predictive Control for Nonprehensile Object Transportation With Obstacle Avoidance on a Mobile ManipulatorIEEE Robotics and Automation Letters, 2023
- FlowFormer: A Transformer Architecture for Optical Flow2022
- Nonprehensile Dynamic Manipulation: A SurveyIEEE Robotics and Automation Letters, 2018