IROS 2018 Workshop: Hands In the Real World
UPDATE: Schedule has been modified. There will be a single poster session for 1 hour in the afternoon.
Connecting End-Effector Design, Sensitivity, and Behavior
October 1, 2018 — 9:00 to 18:30 — Madrid, Spain
Keywords: Dexterous Manipulation, Field Robotics, Mechanisms and Design, Soft Robots, Tactile Sensing, Haptics
Organized by Hannah Stuart (University of California Berkeley), Manuel Giuseppe Catalano (Istituto Italiano di Tecnologia) and Francesca Negrello (Istituto Italiano di Tecnologia)
This workshop is endorsed by the RAS Technical Committee on Robotic Hand Grasping and Manipulation and the RAS Technical Committee on Mechanisms and Design.
This workshop focuses on the field application of robotic end-effectors, lying at the intersection of cutting-edge hardware, behavior and sensitivity. We believe that to overcome traditional end-effector limitations, it is essential to create interconnections between these pillars of robot design.
Experts in controls and sensitivity can inspire new approaches to resilient functionality. Mechanical hardware designers can likewise enable new possibilities in control, through the development of novel hands that can be effectively used outside of the lab. Additionally, what are the trends in performance qualification, and how might this become more efficient and universal. This is a fundamental challenge for hands applied in the field, as contact conditions and environmental circumstances largely influence performance.
The workshop has a strong focus on real applications, therefore it is at the natural intersection between academia and industry. The discussion will involve potential end-users and developers, to identify the applications where soft hands can make the difference, and the needs to expand their use. Young scholars will also be welcomed to contribute through a poster/demo session, and spotlight talks for top contenders. The selected students will have 15 min to present. Given the recent trend of promoting the interactive sessions, we think it is a great opportunity for a student to have a proper speech. This will give them the opportunity to receive feedback from experts and peers in different realms of manipulation research. These scholars are also important in creating new ideas that may elucidate different directions in the field.
Professor, TU Berlin, Robotics and Biology Laboratory
Hand + Sensing + Control = Manipulation – Can We Put the Pieces Together? Much has changed in grasping and manipulation over the past years. Today, robot hands are compliant and underactuated. The environment must not be avoided at all cost but serves as a guide to motion. Point contacts have given way to large contact patches. Static force-closure contact configurations have been replaced with temporally extended, multi-step feedback grasping plans. The consequences of these developments force us to reconsider every component of manipulation. In reconsidering, we must overcome the separation of components and begin to understand integrated manipulation systems, including mechanical design, sensing and perception, and control and planning. In this talk, I will report on the efforts of my lab to do exactly this.
Manuel Giuseppe Catalano
Researcher, Istituto Italiano di Tecnologia (IIT), Soft Robotics for Human Cooperation and Rehabilitation
Connecting hardware & behavior: new opportunities for soft robotic end-effectors: This talk reports on the state of art of artificial hands, discussing some of the most important trends that emerged in the field over the years, and that influenced the modern state of art, suggesting future directions. We review and group the most important application domains of robotic hands, extracting the set of requirements that ultimately led to the use of simplified actuation schemes and soft materials and structures, two trends that clearly emerge from a chronological analysis over the last century. During the talk, we will focus on how the technologies of soft robotics, i.e. the artificial implementation of the principles of impedance adaptation in natural motion control and its role in embodying intelligent manipulation and grasping behaviors, can be used to translate lessons from neuroscience and human motor control into a practically viable and simple artificial hand, the Pisa/IIIT SoftHand. To conclude we will report on the different application domains in which this device found application, and we will show results about novel solutions on the sensing and actuation that are capable to enrich the behaviour of the SoftHand, still maintaing its intrinsic mechanical simplicity and easyness of use.
Professor, Stanford University, Biomimetics and Dexterous Manipulation Laboratory
Hands for a human-scale dexterous climbing quadruped: Although large robots can reach over obstacles, they also present challenges for climbing steep terrain. Mass grows as L^3 but hand and foot areas grow as L^2. To produce rugged end-effectors that can be loaded with hundreds of Newtons while achieving a wide variety of climbing holds, we adopt a combination of soft robotic design and arrays of microspines for high shear stresses on rocky surfaces. This combination introduces its own challenges of combining soft and very hard materials while avoiding undesirable stress concentrations. We use particle jamming to conform to surfaces and anchor the spines. Combined with a tendon-based system to distribute forces, the resulting underactuated hand can achieve a wide variety of holds without relying on feature envelopment.
Associate Professor, Yale University, The Grab Lab
Designing and leveraging soft underactuated hands for dexterous within-hand manipulation: In this talk, I will discuss a range of work being done in my lab to design and control underactuated hands, with a particular focus on dexterous, within-hand manipulation. Other than hand design for various dexterous manipulation modalities, I will discuss work on vision-based control and online learning for model-free within-hand manipulation.
Assistant Professor, Vanderbilt University
Soft Robotic Grippers for Biological Sampling on Deep Reefs: In this presentation, I will introduce our work in the design, bench-top characterization, manufacture, and in situ testing of soft robotic grippers for biological sampling on deep reefs. I will conclude with our recent work on soft robotic sensing that has the potential to support more advanced applications for soft robotic grippers and systems.
Adam Bilodeau, representing Rebecca Kramer
Assistant Professor, Yale University, Faboratory
Proprioceptive and tactile sensing in soft manipulators: During this talk, I will present my group’s progress towards sensory skins – distributed sensors over a surface to provide proprioceptive and tactile feedback. I will discuss the merits of both liquid metal-based resistive sensors and conductive composite-based capacitive sensors in terms of manufacturability and robustness in implementation. I will also show demonstrations of sensory skin prototypes being used for soft robot state estimation and control, including the closed-loop control of a soft gripper designed for pack-and-deploy operations.
Postdoctoral Fellow, Istituto Italiano di Tecnologia (IIT), Soft Robotics for Human Cooperation and Rehabilitation
Soft Robotic Hands for Tough Tasks: Benchmarking robustness: In real-world applications, end effectors are the main media of physical interaction. Consequently, the robotic hand/gripper capability to withstand high contact forces is an essential requirement. In this regard, I will discuss how the design can enhance robustness to impulsive loads and I will propose a method for benchmarking robotic end-effectors under this condition.
Deutsches Zentrum für Luft- und Raumfahrt e.V. (DLR), Institute of Robotics and Mechatronics
Measuring grasp performance: from labs to standardization: Recent years have seen increasing interest in benchmarking grasping and manipulation to foster comparability and reproducibility of results. This talk will discuss different initiatives and procedures for benchmarking robotic hands, encompassing object databases, grasping tests, and future standards for measuring manipulation performance.
Assistant Professor, Scuola Superiore Sant’Anna (Pisa), The BioRobotics Institute
Design of the DETOP neuro-controlled bidirectional artificial hand: The human hand, as well as being one of the principal agents of motor activity, is the chief organ of the fifth sense, touch, and part of our communication system. Its anatomical complexity, richness and variety of sensory receptors combined with its intimate communication link with the brain make the
replacement of a missing hand by a prosthesis one of the big challenges in rehabilitation engineering and applied neuroscience. State of art myoelectric prostheses suffer from a number of limits like: poor/difficult controllability, lack of sensory feedback, poor functionality and cosmesis as well as wearing discomfort (too heavy, too hot or too cold). The speech will summarize the ongoing research concerning the development of a neuro-controlled bidirectional artificial hand in the framework of the EU founded
DETOP project (http://www.detop-project.eu/).
Assistant Professor, University of California Berkeley, Embodied Dexterity Group
Embodying Dexterity for Undersea Worlds: Ocean environments present practical challenges (e.g. leaks, pressure, corrosion) that typically limit hand design. I will discuss how submersion can also provide new possibilities in mechanism sensitivity and control. This case study will set the stage for discussion throughout the workshop.
Associate Professor, University of South Florida, Robot Perception and Action Lab
Grasping for Daily Interactive Manipulation: In this presentation, I will introduce our latest efforts in robotic grasping for daily interactive manipulation tasks. It includes a new data-set that contains interactive force and motion measurements; a manipulation task set that was used in two recent competitions; and a new grasping approach that takes the requirements in interactive manipulation into consideration.
9:15 – 11:15 Session 1
– Embodying dexterity for undersea worlds (Hannah Stuart)
– Connecting hardware & behavior: new opportunities for soft robotic end-effectors (Manuel Catalano)
– Designing and leveraging soft underactuated hands for dexterous within-hand manipulation (Aaron Dollar)
– Towards soft continuous hands for manipulation and interaction (Oliver Brock)
11:15 Coffee Break / Poster Session / Demos
11:30 – 13:00 Session 2
– Soft Robotic Grippers for Biological Sampling (Kevin Galloway)
– Hands for a human-scale dexterous climbing quadruped (Mark Cutkosky)
– Grasping for Daily Interactive Manipulation (Yu Sun)
13:00 – 14:00 Lunch
14:00 – 16:00 Session 3
– Design of the DETOP neuro-controlled bidirectional artificial hand (Marco Controzzi)
– Proprioceptive and tactile sensing in soft manipulators (Adam Bilodeau, filling in for Rebecca Kramer)
– Spotlight talks
16:00 – 17:00 Coffee Break / Poster Session / Demos
17:00 – 18:45 Session 4
– Measuring grasp performance: from labs to standardization (Maximo Roa)
– Soft Robotic Hands for Tough Tasks: Benchmarking robustness (Francesca Negrello)
– Group Discussion
We intend to foster active discussion and participation from all workshop participants: speakers, exhibitors, audience. This is a call for demonstration exhibitors and poster presentations, to allow the workshop’s participants to trigger the discussion.
Separately from the poster/demo session, we have scheduled additional short-form speaking slots that will be awarded to selected contributors. This will be a wonderful opportunity for young scholars to present their work to a prominent crowd. The selected students will have 15 min to present in a podium-style format. Given the recent trend of promoting the interactive sessions, we think it is a great opportunity for a student to have a proper speech, and encourage all to apply.
Upload your information in the survey below. In the PDF file, include a title, abstract, and description of your work or demonstration. Highlight how this work fits into the themes of this particular workshop. Limit 1 page. The initial deadline for entry is September 1, 2018 with notification of acceptance by Sept 10. We have additionally decided to accept entries until Sept 18 with a notification date of Sept 23. Entries after Sept 18 may still be considered, but are not guaranteed.