2nd Workshop on Real-World Physical and Social Human-Robot Interaction in the 2025 IEEE-RAS 24th International Conference on Humanoid Robots (Humanoids), 2nd October 2025, Seoul, South Korea.
The website for the first edition of this workshop at IEEE-Humanoids 2024, featuring proceedings, accepted papers, and a picture gallery, can be accessed at: https://rwhri.github.io/Humanoids2024_workshop/
The workshop will take place in a hybrid mode and available for online participants via Teams: LINK
With the rise of humanoid robots, human-robot interaction (HRI) is undergoing a significant transformation.Unlike traditional robots, humanoids mimic human form and behavior, making natural and high-quality interaction essential. As these robots move from research labs into real-world settings—such as homes, hospitals, and workplaces—HRI must adapt accordingly. Humanoids must not only perceive a wide array of human cues—visual, auditory, tactile, and even emotional—but also respond in socially intelligent, context-aware ways, despite the unpredictability and variability of real environments and human behaviours. Academia and industry approach these challenges differently: the former emphasizes foundational research, while the latter focuses on scalable, deployable solutions. Bridging these perspectives is crucial for the development of humanoids that are not only effective but also accepted in everyday life. This workshop—now in its second edition—aims to bring together leading researchers, practitioners, and industry stakeholders to critically examine the expectations and challenges of interacting with humanoid robots in real-world settings. We will explore the recent advancements in the integration of physical and social HRI for and the design of user-centred interfaces that can flexibly adapt to diverse users and scenarios. Discussion will involve multimodal perception, adaptive human-robot communication, tactile and gestural feedback, sensor fusion, and context-awareness. Building on the success of our inaugural workshop at IEEE Humanoids last year, this year’s event will feature keynote talks by established researchers from academia and industry followed by interactive panel discussions. We will also invite short contributions covering theoretical frameworks, empirical studies, design case studies, and innovative applications, particularly those that highlight the interplay between physical embodiment and social engagement. By fostering dialogue across academia, industry, and physical and social HRI enthusiasts, this workshop seeks to catalyse new collaborations and chart a path towards humanoids capable of seamless, safe, and meaningful human interaction in the real world. To maximize the impact and visibility of the workshop outcomes, we will make all accepted short papers publicly available on this website. Furthermore, we also plan to propose a dedicated special issue in a related RAS journal, inviting authors to extend the short paper contributions or to submit other relevant works inspired by the discussions initiated at the workshop.
This workshop is organized by:
We acknolwedge the financial support by:
The workshop is held in hybrid mode, Please Join the online Meeting: TBD,
If you are attending, please fill up this form alongwith some questions for the panelists in our panel discussions here: TBD
13:20 - 13:30 | Introduction |
13:30 - 14:00 | Keynote 1 – Professor Tetsuya Ogata, Title: Towards Open-Source Foundations in Robotic Control Systems |
14:00 - 14:30 | Keynote 2 – Professor Tamim Asfour, Title: Humanoid Robotics – Engineering Versatile Physical Intelligence |
14:30 - 15:00 | Keynote 3 – Francesco Ferro, CEO, Pal Robotics, Title: 20 Years doing Humanoids |
15:00 - 15:30 | Panel Discussion 1 on “Academia and Industry—Bridging the Gap for Real-World Human-Humanoid Robot Interaction” |
15:30 – 16:00 | Coffee break and Poster Session for Accepted Contributed Papers. |
16:00 - 16:20 | Spotlight Talks from Contributed Papers (See Accepted Papers for more) |
16:20 - 16:45 | Keynote 5 – Assistant Professor Maria Pozzi, Title: User-centered human-robot augmentation |
16:45 - 17:05 | Keynote 6 – Postdoctoral Researcher Marta Lagomarsino, Title: Toward Adaptive and Personalised Robot Behaviour via Human State Monitoring and Expert Feedback |
17:05 - 17:10 | Closing Remarks |
Meet our esteemed speakers from academia and industry.
Professor, Karlsruhe Institute of Technology, Germany
Panelist for Panel Discussion
Presentation Title: Humanoid Robotics – Engineering Versatile Physical Intelligence
Humanoid robots represent one of the most ambitious goals in robotics: creating machines with versatile physical intelligence to operate seamlessly in human-centered environments. The talk discusses the current state of humanoid robotics, focusing on the engineering challenges of developing embodied holistic systems that integrate AI and mechatronics to perform complex manipulation tasks and interact naturally with humans. Examples from research and emerging applications will demonstrate current progress and highlight the requirements for achieving truly versatile physical intelligence in humanoid robot systems acting in real-world environments. The talk will address fundamental questions about realistic deployment timelines for and identifies critical research directions needed for the transition from laboratories to homes and workplaces in the coming years.
Professor Waseda University, Joint Appointed Fellow, AIST, Japan
Panelist for Panel Discussion
Presentation Title: Towards Open-Source Foundations in Robotic Control Systems
Presentation Abstract: In recent years, there has been an intensified global effort, particularly in the United States and China, to develop foundational robot models capable of End-to-End robot control. These models hold the promise of accelerating robotic advancements and significantly broadening their applications across various domains. However, while the models themselves are becoming increasingly accessible, key utilization data and development methodologies remain largely undisclosed, posing challenges to their practical application. In this presentation, I will introduce the ecosystem for open foundational robot model development within the AI Robot Association (AIRoA), where I serve as Chairman, and discuss the prospects of these models in advancing robotics research and application.
Panelist for Panel Discussion
Presentation:20 Years doing Humanoids
Assistant Professor, University of Siena, Italy
Presentation Title: User-centered human-robot augmentation
Presentation Abstract: The idea of augmenting human manipulation capabilities through supernumerary robotic limbs which move in coordination with the biological ones is potentially disruptive in the field of human-robot interaction. It finds its most compelling application in assisting people with disabilities and requires a deep connection between the human and the robot. In this context, a user-centered perspective is essential to identify the diverse needs of end-users and ensure the envisaged level of human-robot sensorimotor integration through suitably designed interfaces
Postdoctoral Researcher, Italian Institute of Technology, Italy
Presentation: Toward Adaptive and Personalised Robot Behaviour via Human State Monitoring and Expert Feedback
Presentation Abstract: Robotics and AI have reached remarkable capabilities in complex tasks and environments. However, meaningful human engagement with these technologies remains limited, as robots often struggle to understand and personalise interactions to human needs and intentions. This talk presents methods to infer human psycho-physical states and preferences by observing movement patterns during interaction, and shows how this information can be exploited to adapt robot behavior for workload mitigation and seamless collaboration. It will discuss the intuitive transfer of human skills to robots by capturing hand movements and object interactions in a video demonstration and refining the robot plans via natural language inputs. Together, these directions point toward more personalised, adaptive, and intuitive human-robot collaboration.
Professor, Tokyo University of Science, Japan
Panelist for Panel Discussion
ISFP Researcher, Centre Inria de l'Université de Lorraine & Loria, Nancy, France
Panelist for Panel Discussion
Authors: Jin Cheng, Dongho Kang, Gabriele Fadini, Guanya Shi, Stelian Coros
Loco-manipulation, physical interaction of various objects that is concurrently coordinated with locomotion, remains a major challenge for legged robots due to the need for both precise end-effector control and robustness to unmodeled dynamics. While model-based controllers provide precise planning via online optimization, they are limited by model inaccuracies. In contrast, learning-based methods offer robustness, but they struggle with precise modulation of interaction forces.
We introduce RAMBO, a hybrid framework that integrates model-based whole-body control within a feedback policy trained with reinforcement learning. The model-based module generates feedforward torques by solving a quadratic program, while the policy provides feedback corrective terms to enhance robustness. We validate our framework on a quadruped robot across a diverse set of real-world loco-manipulation tasks, such as pushing a shopping cart, balancing a plate, and holding soft objects, in both quadrupedal and bipedal walking. Our experiments demonstrate that RAMBO enables precise manipulation capabilities while achieving robust and dynamic locomotion.
Authors: Damiano Gasperini, Luca Muratore, Nikos Tsagarakis
The advancement of autonomous robots still needs a comprehensive robot awareness framework for the execution of human-robot tasks, enabling the generation of autonomous behaviors and responses in dynamic environments. Situation awareness aims to enhance robot’s autonomy and adaptation during task execution by effectively combining key capabilities, such as reasoning, planning, projection of action’s effects into future states, as well as perception and comprehension of the surroundings.
In this work, we propose a novel AI-based framework for semantic-driven task awareness that addresses the intricate challenges of perceiving, navigating, and manipulating dynamic environments. Our objective of achieving effective robot awareness is based on the perceived environment semantics and on the combined use of online planning, reasoning, and monitoring, while also allowing recovery from task-level failures. The presented method is designed as a set of online modules for (a) collecting and analyzing sensors’ data as well as updating the world model description, (b) real-time decision-making and task-state monitoring, and (c) execution of each action. Being highly modular and configurable to assorted robotic systems, the proposed framework aims to adapt to diverse robotic platforms and tasks. In this work, a preliminary description of the framework comes along with demonstrative human-robot tasks to showcase its capabilities on both the CENTAURO robot and on a custom 6 DoF manipulator.
Authors: Vignesh Prasad, Zixuan Fang, Georgia Chalvatzaki
To enable seamless Human-Robot Interaction (HRI), a robot must react to a human in a timely manner while accurately maintaining physical proximity with the human in the context of the task at hand. Recent approaches have shown competitive performance in learning multimodal behaviours for Human-Robot Interactions from demonstrations. In the context of learning-based approaches, ensuring fast inference can enable robots to act more reactively to the human interaction partner, thereby improving the perceived interactiveness of the robot.
We propose InteractionFlow, a novel method that models a joint distribution between the human and the robot actions using recent advancements in Imitation Learning using Conditional Flow Matching to enable learning fast, reactive, interactive behaviours from demonstrations. We find that using the generative abilities of Flow Matching enables InteractionFlow to achieve state-of-the-art results on example human-robot collaboration tasks in simulation, which require a high degree of precision while ensuring a seamless and reactive interaction.
We invite short paper submissions presenting late-breaking results, novel ideas, design case studies, or empirical findings on real-world physical and social human-robot interaction with humanoids. Accepted authors will receive ~5 minutes for spotlight presentations and participate in a 40-minute interactive poster session, engaging with an expected audience of 50–60 participants, including leading experts from academia and industry. Submit your work for a valuable opportunity to share, discuss, and collaborate on the key challenges and advances shaping the future of Human-Humanoids Robot Interaction.
Date: October 2, 2025 (Full-day workshop)
Location: Hybrid (Seoul, South Korea and online), as part of The 2025 IEEE-RAS International Conference on Humanoid Robots, IEEE-Humanoids 2025.
Submission Instruction: Email your contributions to: whsop.realworld.hri@gmail.com
Important: Special Session in a IEEE-RAS Journal! We are currently in the process of submitting a proposal for a dedicated special issue in a related IEEE-Robotics and Automation Society journal (e.g. RA-L), focusing on advances in physical and social human-humanoid interaction, multimodal perception, adaptive communication, and real-world applications with humanoid robots, reflecting the key themes of the workshop. All workshop contributions will be invited to submit extended versions of their work for consideration in this special issue. If you are a researcher interested in contributing to this special issue, please contact us for more information.
Contact for submissions: paragk@kth.se, davide.torielli@iit.it.
All deadlines are at 23:59 Anywhere on Earth time.
Manuscripts should be written in English and will undergo a single-blind review by the organizing committee and selected researchers from the field. The length should be 2-4 pages excluding references. We welcome contributions that include work in progress, preliminary results, technical reports, case studies, surveys, and state-of-the-art research. Position papers are also welcome and should be at least 2 pages excluding references. These can be research project proposals or plans without results. Authors must use the Humanoids templates provided, formatted for US Letter. The templates can be downloaded below.
Organizers:
PostDoctoral Researcher, Italian Institute of Technology, Italy: davide.torielli@iit.it
Lead Researcher Engineer, Finger Vision Inc., Japan: angela.faragaso@fingervision.biz
Principal Investigator, Italian Institute of Technology, Italy: Nikos.tsagarakis@iit.it
We acknolwedge the financial support by: