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

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Liu, Zeyi, Chi, Cheng, Cousineau, Eric, Kuppuswamy, Naveen, Burchfiel, Benjamin, Song, Shuran
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910681147637760
author Liu, Zeyi
Chi, Cheng
Cousineau, Eric
Kuppuswamy, Naveen
Burchfiel, Benjamin
Song, Shuran
author_facet Liu, Zeyi
Chi, Cheng
Cousineau, Eric
Kuppuswamy, Naveen
Burchfiel, Benjamin
Song, Shuran
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2406_19464
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle ManiWAV: Learning Robot Manipulation from In-the-Wild Audio-Visual Data
Liu, Zeyi
Chi, Cheng
Cousineau, Eric
Kuppuswamy, Naveen
Burchfiel, Benjamin
Song, Shuran
Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
Sound
Audio and Speech Processing
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.
title ManiWAV: Learning Robot Manipulation from In-the-Wild Audio-Visual Data
topic Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
Sound
Audio and Speech Processing
url https://arxiv.org/abs/2406.19464