Contents
| | | 18 Publications and References | Contents |
Copyright notice: this material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.
All of the publication below are available at http://grasping.cs.columbia.edu.
For an introduction to GraspIt!, the most complete overview of its core features is in:
We recommend starting with that paper for the best introduction to the system. Most of the papers below address individual features of the simulator, you can read those that are relevant to the particular project you are working on. The list of publication is presented in chronological order, from oldest to newest. For each publication, we also provide a short description of the parts of GraspIt! that it is most relevant for.
Introductory theory on the grasp quality metrics used by GraspIt!. Discussed topics such as the Grasp Wrench Space, L1 and LInf norms, epsilon and volume quality metrics, etc.
Extremely detailed presentation of the GraspIt! core. The presentation is mostly from a theoretical and research standpoint, but also covers a number of practical implementation issues.
Application of GraspIt! to execute a grasping task with a real robot. A real-life object is tracked using a camera, its position is replicated in GraspIt! where a grasp is planned using a virtual Barrett hand. The grasp is then executed using a real Barrett hand.
Detailed discussion of the Primitive-based grasp planner.
Presents the theoretical framework between the dynamics engine in GraspIt!. Covers topics such as time step integration, formulation of contact and joint constraints as Linear Complementarity constraints, etc. Shows how the full Linear Complementarity problem is assembled and solved at each time step of the dynamic engine. A must-read for understanding GraspIt! dynamics.
Proposed the use of GraspIt! to generate large amounts labeled grasping data that can be used to apply machine learning algorithms. This code is not included in the current GraspIt! distribution.
Discusses the Soft Finger contact as implemented in GraspIt!, covering the analytical surface approximation, soft finger grasp wrench space and formulation as linear complementarity constraints.
Proposes an automatic method of decomposing an object into primitives (superquadrics) to fully automate the task of primitive-based grasp planning. This code is not included in the current GraspIt! distribution.
Introduces the eigengrasp concept and grasp planning in eigengrasp space as an optimization problem solved through Simulated Annealing. This is the recommended starting point if you are interested in eigengrasps.
Presents an application of eigengrasp planning for on-line interaction with a human user. This is the theory behind the OnLinePlanner class included in the distribution.
These two papers show how GraspIt! can be used to generate a huge database of labeled grasp data, and how this database can be used for data-driven grasp planning algorithms. The database is publicly available. GraspIt! also provides an interface for accessing and using this information; see the Columbia Grasp Database chapter in this manual.
Presents one use of the GraspIt Grasp Force Optimization (GFO) engine and introduces the mathematical framework used by this engine.
Copyright (C) 2002-2009 Columbia University
| | | 18 Publications and References | Contents |