Reproducible Benchmarking of

Robotic Grasping and Manipulation:

From Advanced AI to Generalized Humanoid Intelligence

______________

1 October 2026, 13:30 - 17:30 (local time)

2026 IEEE/RSJ International Conference on Intelligent Robots and Systems

PITTSBURGH, PA, USA

Advances in robotics, perception, and AI are raising expectations for robots to achieve high-level capabilities that enable them to autonomously, accurately, and robustly perform complex manipulation tasks. These include autonomous assembly and disassembly, in-hand manipulation, heterogeneous object picking in cluttered environments, and human–robot collaboration. Despite significant research efforts over the past decades, many of these crucial capabilities remain far from solved. Most existing solutions are demonstrated through custom pipelines and validation procedures developed within individual laboratories, often relying on different sensing modalities, workflows, and evaluation criteria. As a result, it is difficult to fairly compare approaches or reproduce results across research groups.

Benchmarking, therefore, plays a critical role in enabling reproducibility, transparent evaluation, and meaningful comparison of different methods, ultimately accelerating progress in challenging robotic manipulation problems. This workshop aims to bring together researchers, practitioners, and stakeholders from academia, industry, and standardization organizations to discuss the current status, challenges, and future directions of reproducible benchmarking in autonomous robotic grasping and manipulation.

Particular attention will be given to the growing interaction between research innovation and practical engineering solutions needed to develop efficient, robust, and trustworthy robotic systems that operate reliably in real-world environments outside the laboratory. The lessons learned from the workshop will provide insights into the evolution of grasping and manipulation research and will feed the development and deployment of humanoids and advanced artificial intelligence systems.

Sponsor

$300 cash prize for the Best Extended/Poster award
Banquet dinner for the invited speakers

Complimentary IROS 2026 Workshops

Reproducing Robot Manipulation: Developing Guidelines to Improve Accessibility and Reproducibility of Open-Source
September 27, 1:00 - 5:30 pm (Local time)

Invited Speakers

Kris Hauser
University of Illinois

and

Samsung Research America

Rika Antonova

University of Cambridge

Natasha Kholgade Banerjee

Wright State University

Kostas Bekris

Rutgers University

and

Amazon.com Inc

Aswin Ramachandran

ETH Zürich

Representatives of the teams from the tracks of the 11th Robotic Grasping and Manipulation Competition at ICRA 2026 in Vienna

Qiyan Ke

Winning team
Picking in Clutter

Frano Petric

Team representative
Human-to-robot handover

Mingwei Ma

Team representative
Cloud robotics

List of Abstracts (published works) and Extended Abstracts (unpublished works)

See https://sites.google.com/view/rgmcw2026/contributions/accepted-works for details and PDFs.

Award

$300 for the Best Extended Abstract
Sponsored by Nokov Motion Capture

Abstracts

What Are We Actually Benchmarking in Robot Manipulation? [Best Abstract]

Tianchong Jiang, Xiangshan Tan, Samuel Wheeler, Luzhe Sun, Tewodros W. Ayalew, Matthew Walter

Conference on Robot Learning (CoRL), 2026
ArXiv: https://doi.org/10.48550/arXiv.2606.04233

PhyPush: One Push is All You Need for Sensorless Physical Property Estimation with Physics-Guided Transformers

Koyo Fujii, Luis Figueredo, Praminda Caleb-Solly, Ivan Boschi, Edoardo Ida’, Marco Carricato, Aly Magassouba

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2026

ArXiv: https://doi.org/10.48550/arXiv.2605.26284

Extended Abstracts

Object Pose Is Not Enough: Combined Finger--Object Task Descriptions for In-Hand Manipulation Benchmarking [Best Extended Abstract]

Shambhuraj A. Mane, Andrew S. Morgan, Berk Calli

Clutter Metrics and Software Tools to Evaluate Multi-Object Scenes for Benchmarking Robot Grasping and Manipulation

John Brann, Peter Gavriel, Adam Norton, Brian Flynn, Holly Yanco

GhostGrasp: Grasping Transparent Objects in Darkness from Depth-Failure Evidence

Runfeng Ling

MATE: Mask-Guided Adaptive Object Template Estimation and Tactile-Confirmed Execution for Human-to-Robot Handover

Seungheon Jeong∗, Sebin Lee∗, Minji Kim, Junsu Kim, Kyungdon Joo

HiveBoard: An Open, Modular, 3D-Printed Benchmark of Industrial Mechanisms for Robotic and Prosthetic Manipulation

Ricardo V. Godoy, Enzo F. de Souza, Rudy De-Xin de Lange, Juliano Negri, Joao Aires Marsicano, Victor I. van Halst, Aravind Elanjimattathil Vijayan, Gianluca Capezzuto,
Matheus P. Angarola, Felipe A. G. Tommaselli, Giuseppe Milazzo, Rafael R. Baptista, Meiko Adriana van Berge, Gustavo J. G. Lahr, Lucas Ferrari Gerez, Antonio Bicchi, Marcelo Becker

Confidence-Aware Multimodal Language-Guided Grasping on a Low- Cost Robotic Manipulator

Rishabh Gupta, Krishiv Garg, Sanchit Kumar Shamra, Priya Kansal, Sushil Raut, Sachin Kansal

Planar translation capability of two-fingered robotic hands using human studies

John Morrow, Ravi Balasubramanian, Cindy Grimm

Industrial Dexterity Benchmark (IDB): Open-Source Benchmarking Boards for Datacenter Cable Management, Automotive Cable Harness, and Gearbox Assembly Tasks

Honglu He

Program (local time)

13:30 - 13:40 Welcome and introduction

Salvatore D'Avella/Yu Sun

[Session 1] Human-to-Robot Handover

13:40 - 13:55 Enhancing AI for Automation and Robotics Through Awareness of Human-Object Interactions

Natasha K. Banerjee, Wright State University (USA)

13:55 - 14:05 Soft gripper and soft AI for human-to-robot handover

Frano Petric, University of Zagreb

Representative of LARICS Gripsters team at 2026 ICRA RGMC

14:05- 14:10 Q&A

[Session 2] Grasping Heterogeneous Objects In Cluttered Environments

14:10 - 14:25 How Will We Know Robot Grasping Is Solved?

Kostas Bekris, Rutgers University and Amazon Robotics

14:25 - 14:35 A Multi-Sensor Multi-End-Effector Robotic System for Ordered Picking in Cluttered Bins

Qiyan Ke, Tsinghua University

Representative of the winning team for the 2026 ICRA RGMC - Picking in Clutter track

14:35 - 14:40 Q&A

[Session 3] Cloud Robotics

14:40 - 14:55 Teaching Robots to Master Contact-Rich Adversarial Board Games: KLASK

Aswin Karthik Ramachandran Venkatapathy, ETH Zürich

14:55 - 15:05 From Planar Pushing to Rope Shaping: Generalizable Robotic Manipulation on CloudGripper

Mingwei Ma, Institution: South China University of Technology

Representative of the winning team for the 2026 ICRA RGMC - Cloud Robotics track

15:05 - 15:10 Q&A

[Session 4] Best abstract and extended abstract

15:10 - 15:20 Object Pose Is Not Enough: Combined Finger--Object Task Descriptions for In-Hand Manipulation Benchmarking [Best Extended Abstract]

Shambhuraj Mane, Worcester Polytechnic Institute

15:20 - 15:30 What Are We Actually Benchmarking in Robot Manipulation? [Best Abstract]

Tianchong Jiang, Toyota Technological Institute at Chicago

15:30 - 15:35 Q&A

15:30 - 16:00 Poster Session

[Session 5] Generative AI and Humanoids

16:00 - 16:15 Bridging the Lab to Production Gap in Manufacturing Skills

Kris Hauser, Samsung Research America

16:15 - 16:30 Co-designing the Future of Robotics: Benchmarks, Milestones, and Open Challenges

Rika Antonova, University of Cambridge

16:30 - 16:40 Q&A

16:40 - 16:45 Sponsor Talk: Amy Zhang from Nokov Motion Capture

16:45 - 17:15 Panel Discussion

Salvatore D'Avella/Yu Sun

17:30 End

Organizers

Salvatore D'Avella

Assistant Professor

Alessio Xompero

Independent Researcher

Changjae Oh

Lecturer

Andrea Cavallaro

Full Professor

Florian T. Pokorny

Professor

Omar Aboul-Enein

Computer Scientist

Yu Sun

Full Professor

Contact

For further questions please send an email to salvatore.davella@santannapisa.it

The Microsoft CMT service was used for managing the peer-reviewing process for this conference. This service was provided for free by Microsoft and they bore all expenses, including costs for Azure cloud services as well as for software development and support.

Page updated

Google Sites

Report abuse