Data-efficient Model Learning and Prediction for Contact-rich Manipulation Tasks

TL;DR

This paper introduces a data-efficient hybrid modeling approach using Gaussian processes for contact-rich manipulation tasks, improving long-term prediction accuracy in low-data scenarios within model-based reinforcement learning.

Contribution

It proposes a novel hybrid model structure that explicitly handles discontinuous dynamics and leverages Gaussian process uncertainty for enhanced data efficiency.

Findings

Outperforms baseline methods in low-data regimes

Demonstrates effective long-term prediction in contact-rich tasks

Shows advantages on a 7-DOF robot and moving block task

Abstract

In this letter, we investigate learning forward dynamics models and multi-step prediction of state variables (long-term prediction) for contact-rich manipulation. The problems are formulated in the context of model-based reinforcement learning (MBRL). We focus on two aspects-discontinuous dynamics and data-efficiency-both of which are important in the identified scope and pose significant challenges to State-of-the-Art methods. We contribute to closing this gap by proposing a method that explicitly adopts a specific hybrid structure for the model while leveraging the uncertainty representation and data-efficiency of Gaussian process. Our experiments on an illustrative moving block task and a 7-DOF robot demonstrate a clear advantage when compared to popular baselines in low data regimes.