Robotic Grasping of Fully-Occluded Objects using RF Perception
We present the design, implementation, and evaluation of RF-Grasp, a robotic system that can grasp fully-occluded objects in unknown and unstructured environments. Unlike prior systems that are constrained by the line-of-sight perception of vision and infrared sensors, RF-Grasp employs RF (Radio...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2022
|
| Online Access: | https://hdl.handle.net/1721.1/146572 |
| _version_ | 1826212393737256960 |
|---|---|
| author | Boroushaki, Tara Leng, Junshan Clester, Ian Rodriguez, Alberto Adib, Fadel |
| author2 | Massachusetts Institute of Technology. Media Laboratory |
| author_facet | Massachusetts Institute of Technology. Media Laboratory Boroushaki, Tara Leng, Junshan Clester, Ian Rodriguez, Alberto Adib, Fadel |
| author_sort | Boroushaki, Tara |
| collection | MIT |
| description | We present the design, implementation, and evaluation of RF-Grasp, a robotic
system that can grasp fully-occluded objects in unknown and unstructured
environments. Unlike prior systems that are constrained by the line-of-sight
perception of vision and infrared sensors, RF-Grasp employs RF (Radio
Frequency) perception to identify and locate target objects through occlusions,
and perform efficient exploration and complex manipulation tasks in
non-line-of-sight settings.
RF-Grasp relies on an eye-in-hand camera and batteryless RFID tags attached
to objects of interest. It introduces two main innovations: (1) an RF-visual
servoing controller that uses the RFID's location to selectively explore the
environment and plan an efficient trajectory toward an occluded target, and (2)
an RF-visual deep reinforcement learning network that can learn and execute
efficient, complex policies for decluttering and grasping.
We implemented and evaluated an end-to-end physical prototype of RF-Grasp. We
demonstrate it improves success rate and efficiency by up to 40-50% over a
state-of-the-art baseline. We also demonstrate RF-Grasp in novel tasks such
mechanical search of fully-occluded objects behind obstacles, opening up new
possibilities for robotic manipulation. Qualitative results (videos) available
at rfgrasp.media.mit.edu |
| first_indexed | 2024-09-23T15:20:42Z |
| format | Article |
| id | mit-1721.1/146572 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T15:20:42Z |
| publishDate | 2022 |
| publisher | Institute of Electrical and Electronics Engineers (IEEE) |
| record_format | dspace |
| spelling | mit-1721.1/1465722023-02-16T16:17:25Z Robotic Grasping of Fully-Occluded Objects using RF Perception Boroushaki, Tara Leng, Junshan Clester, Ian Rodriguez, Alberto Adib, Fadel Massachusetts Institute of Technology. Media Laboratory Massachusetts Institute of Technology. Department of Mechanical Engineering Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science We present the design, implementation, and evaluation of RF-Grasp, a robotic system that can grasp fully-occluded objects in unknown and unstructured environments. Unlike prior systems that are constrained by the line-of-sight perception of vision and infrared sensors, RF-Grasp employs RF (Radio Frequency) perception to identify and locate target objects through occlusions, and perform efficient exploration and complex manipulation tasks in non-line-of-sight settings. RF-Grasp relies on an eye-in-hand camera and batteryless RFID tags attached to objects of interest. It introduces two main innovations: (1) an RF-visual servoing controller that uses the RFID's location to selectively explore the environment and plan an efficient trajectory toward an occluded target, and (2) an RF-visual deep reinforcement learning network that can learn and execute efficient, complex policies for decluttering and grasping. We implemented and evaluated an end-to-end physical prototype of RF-Grasp. We demonstrate it improves success rate and efficiency by up to 40-50% over a state-of-the-art baseline. We also demonstrate RF-Grasp in novel tasks such mechanical search of fully-occluded objects behind obstacles, opening up new possibilities for robotic manipulation. Qualitative results (videos) available at rfgrasp.media.mit.edu 2022-11-21T19:42:31Z 2022-11-21T19:42:31Z 2021 2022-11-21T18:31:03Z Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/146572 Boroushaki, Tara, Leng, Junshan, Clester, Ian, Rodriguez, Alberto and Adib, Fadel. 2021. "Robotic Grasping of Fully-Occluded Objects using RF Perception." 2021 IEEE International Conference on Robotics and Automation (ICRA). en 10.1109/ICRA48506.2021.9560956 2021 IEEE International Conference on Robotics and Automation (ICRA) Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Institute of Electrical and Electronics Engineers (IEEE) MIT web domain |
| spellingShingle | Boroushaki, Tara Leng, Junshan Clester, Ian Rodriguez, Alberto Adib, Fadel Robotic Grasping of Fully-Occluded Objects using RF Perception |
| title | Robotic Grasping of Fully-Occluded Objects using RF Perception |
| title_full | Robotic Grasping of Fully-Occluded Objects using RF Perception |
| title_fullStr | Robotic Grasping of Fully-Occluded Objects using RF Perception |
| title_full_unstemmed | Robotic Grasping of Fully-Occluded Objects using RF Perception |
| title_short | Robotic Grasping of Fully-Occluded Objects using RF Perception |
| title_sort | robotic grasping of fully occluded objects using rf perception |
| url | https://hdl.handle.net/1721.1/146572 |
| work_keys_str_mv | AT boroushakitara roboticgraspingoffullyoccludedobjectsusingrfperception AT lengjunshan roboticgraspingoffullyoccludedobjectsusingrfperception AT clesterian roboticgraspingoffullyoccludedobjectsusingrfperception AT rodriguezalberto roboticgraspingoffullyoccludedobjectsusingrfperception AT adibfadel roboticgraspingoffullyoccludedobjectsusingrfperception |