ManiWAV: Learning Robot Manipulation from In-the-Wild Audio-Visual Data

TL;DR

ManiWAV introduces an innovative 'ear-in-hand' device for collecting synchronized audio-visual demonstrations in natural environments, enabling robots to learn contact-rich manipulation skills more effectively from diverse, real-world data.

Contribution

The paper presents a novel data collection system and policy learning framework that leverage in-the-wild audio-visual data for robot manipulation, expanding beyond traditional teleoperated methods.

Findings

Successfully learned manipulation policies from in-the-wild demonstrations.

Demonstrated generalization to unseen environments.

Achieved effective sensing of contact events and object properties.

Abstract

Audio signals provide rich information for the robot interaction and object properties through contact. This information can surprisingly ease the learning of contact-rich robot manipulation skills, especially when the visual information alone is ambiguous or incomplete. However, the usage of audio data in robot manipulation has been constrained to teleoperated demonstrations collected by either attaching a microphone to the robot or object, which significantly limits its usage in robot learning pipelines. In this work, we introduce ManiWAV: an 'ear-in-hand' data collection device to collect in-the-wild human demonstrations with synchronous audio and visual feedback, and a corresponding policy interface to learn robot manipulation policy directly from the demonstrations. We demonstrate the capabilities of our system through four contact-rich manipulation tasks that require either passively sensing the contact events and modes, or actively sensing the object surface materials and states. In addition, we show that our system can generalize to unseen in-the-wild environments by learning from diverse in-the-wild human demonstrations.