Incremental Few-Shot Adaptation for Non-Prehensile Object Manipulation using Parallelizable Physics Simulators

TL;DR

This paper introduces an incremental adaptation method for non-prehensile object manipulation that uses parallelizable physics simulators and model-predictive control to improve robot performance with minimal examples.

Contribution

It presents a novel approach combining parallelizable physics simulation and sampling-based optimization for incremental model adaptation in robotic manipulation.

Findings

Effective in simulation for object pushing tasks

Successfully applied to real robot experiments

Improves manipulation performance with few examples

Abstract

Few-shot adaptation is an important capability for intelligent robots that perform tasks in open-world settings such as everyday environments or flexible production. In this paper, we propose a novel approach for non-prehensile manipulation which incrementally adapts a physics-based dynamics model for model-predictive control (MPC). The model prediction is aligned with a few examples of robot-object interactions collected with the MPC. This is achieved by using a parallelizable rigid-body physics simulation as dynamic world model and sampling-based optimization of the model parameters. In turn, the optimized dynamics model can be used for MPC using efficient sampling-based optimization. We evaluate our few-shot adaptation approach in object pushing experiments in simulation and with a real robot.