# Learning Hybrid Object Kinematics for Efficient Hierarchical Planning Under Uncertainty

arXiv: https://arxiv.org/abs/1907.09014 · HTML: https://mlsanity.com/paper/1907.09014/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Jul 2019.

## Built on by (most cited first, 9 of 9)

- [UMPNet: Universal Manipulation Policy Network for Articulated Objects](https://mlsanity.com/paper/2109.05668/) (2021)
- [End-to-End Affordance Learning for Robotic Manipulation](https://mlsanity.com/paper/2209.12941/) (2022)
- [CLA-NeRF: Category-Level Articulated Neural Radiance Field](https://mlsanity.com/paper/2202.00181/) (2022)
- [Category-Independent Articulated Object Tracking with Factor Graphs](https://mlsanity.com/paper/2205.03721/) (2022)
- [Distributional Depth-Based Estimation of Object Articulation Models](https://mlsanity.com/paper/2108.05875/) (2021)
- [Learning Category-Level Generalizable Object Manipulation Policy via Generative Adversarial Self-Imitation Learning from Demonstrations](https://mlsanity.com/paper/2203.02107/) (2022)
- [Learning to Infer Kinematic Hierarchies for Novel Object Instances](https://mlsanity.com/paper/2110.07911/) (2021)
- [Self-Supervised Learning of Action Affordances as Interaction Modes](https://mlsanity.com/paper/2305.17565/) (2023)
- [Learning the Generalizable Manipulation Skills on Soft-body Tasks via Guided Self-attention Behavior Cloning Policy](https://mlsanity.com/paper/2410.05756/) (2024)

## Abstract (arXiv)

Sudden changes in the dynamics of robotic tasks, such as contact with an object or the latching of a door, are often viewed as inconvenient discontinuities that make manipulation difficult. However, when these transitions are well-understood, they can be leveraged to reduce uncertainty or aid manipulation---for example, wiggling a screw to determine if it is fully inserted or not. Current model-free reinforcement learning approaches require large amounts of data to learn to leverage such dynamics, scale poorly as problem complexity grows, and do not transfer well to significantly different problems. By contrast, hierarchical POMDP planning-based methods scale well via plan decomposition, work well on novel problems, and directly consider uncertainty, but often rely on precise hand-specified models and task decompositions. To combine the advantages of these opposing paradigms, we propose a new method, MICAH, which given unsegmented data of an object's motion under applied actions, (1) detects changepoints in the object motion model using action-conditional inference, (2) estimates the individual local motion models with their parameters, and (3) converts them into a hybrid automaton that is compatible with hierarchical POMDP planning. We show that model learning under MICAH is more accurate and robust to noise than prior approaches. Further, we combine MICAH with a hierarchical POMDP planner to demonstrate that the learned models are rich enough to be used for performing manipulation tasks under uncertainty that require the objects to be used in novel ways not encountered during training.