Self-Adapting Recurrent Models for Object Pushing from Learning in Simulation

TL;DR

This paper introduces a self-adapting recurrent model for object pushing that is trained in simulation and adapts quickly to real-world dynamics, improving robotic pushing performance with minimal real data.

Contribution

It presents a novel recurrent model predictive control framework that adapts to real object dynamics using simulation-trained models and domain randomization.

Findings

The model adapts to real object dynamics within a few steps.

The proposed RMPPI algorithm improves path planning with recurrent models.

Experiments show effective pushing on a UR5 robot platform.

Abstract

Planar pushing remains a challenging research topic, where building the dynamic model of the interaction is the core issue. Even an accurate analytical dynamic model is inherently unstable because physics parameters such as inertia and friction can only be approximated. Data-driven models usually rely on large amounts of training data, but data collection is time consuming when working with real robots. In this paper, we collect all training data in a physics simulator and build an LSTM-based model to fit the pushing dynamics. Domain Randomization is applied to capture the pushing trajectories of a generalized class of objects. When executed on the real robot, the trained recursive model adapts to the tracked object's real dynamics within a few steps. We propose the algorithm \emph{Recurrent} Model Predictive Path Integral (RMPPI) as a variation of the original MPPI approach, employing state-dependent recurrent models. As a comparison, we also train a Deep Deterministic Policy Gradient (DDPG) network as a model-free baseline, which is also used as the action generator in the data collection phase. During policy training, Hindsight Experience Replay is used to improve exploration efficiency. Pushing experiments on our UR5 platform demonstrate the model's adaptability and the effectiveness of the proposed framework.