Research in robot grasping has experienced a massive shift in recent years. Early research into the field focused on synthesizing grasps analytically, using precise and accurate models of the objects to be grasped. Although these approaches are well-suited for grasping objects in industrial scenarios, such as on an assembly line, they are not easily applicable to unstructured environments like households or hospitals.
During the past fifteen years, grasping research has shifted to methods that can cope with novel objects and partial/noisy perceptual data. The challenges posed by grasping in these situations have led to the development of a wide range of new algorithms for selecting, executing, and evaluating grasps. In parallel to the progress made on the software level, new robot hands and sensors have also been designed for operating in everyday environments.
The first aim of the workshop is to report progress in the large variety of grasping and manipulation methods. The second aim of the workshop is to discuss the increasing difficulty in comparing the practical applicability of all the new algorithms and hardware that appear each month. Comparing different methodologies in a fair manner is difficult. The success of a grasp depends on many factors, such as the type of robot hand/gripper, the immediate environment around the object, and the properties of the objects being grasped. Different methods might also assume different prior knowledge regarding the object, or rely on the robot being equipped with certain sensors. The lack of suitable means of comparizon has become a serious concern. The workshop will serve as a platform to discuss the role of benchmarking in grasping and manipulation. In particular, we will animate a discussion on the pros, cons, benefits and pitfalls of benchmarking via a periodic community-organized grasping/manipulation challenge.
ICRA 2014 is the ideal venue for this workshop. Benchmarking can only be done if a large number groups adopt the same practices. Hence, it is important that the topic of benchmarking is discussed on the international level, in a venue that attracts a major fraction of the community.
Format: Full-day workshop.
Main organizer:
Co-organizers:
The workshop will run for a full day and consist of a mix of presentations by invited speakers and selected contributions, as well as poster sessions. We have received confirmation from eight top-level speakers, listed below. Abstracts describing novel techniques and systems directly relevant to the workshop theme will be solicited from appropriate communities. The workshop proceedings will be published online.
Submissions will be extended abstracts of 1-2 pages in length. Each paper will be single-blind peer-reviewed by two reviewers for relevance and novelty. The accepted papers will be presented during poster sessions to allow for more discussion and for presenters to obtain additional feedback on their work. Exceptional contributions will also be given the opportunity to present their work infront of the entire workshop audience as a presentation.
The workshop will conclude with a panel discussion involving speakers and audience. The organizers plan to write a position paper based on the discussions and conclusions of the workshop, in the form of a roadmap for the specific goal of organizing a grasping/manipulation challenge.
The aims of this workshop are twofold. First, the workshop will bring together researchers from the grasping and manipulation communities, to exchange ideas and results towards the goal of developing autonomous manipulation systems. Second, the workshop will serve as a platform to openly discuss the organization of a grasping challenge.
The workshop is aimed at both students and experienced researchers. The presentation of the state of the art, and the discussion of best practices in benchmarking, are both relevant to every member of the community.
All of the speakers below have confirmed their participation.
We will solicit submissions in the broad area of grasping and manipulation. Topics of interest addressed by this workshop include, but are not limited to:
We will encourage submissions related to benchmarking, and on the benefits/pitfalls of having an open grasping challenge.
Grasping and manipulation have been discussed at the following recent events: