| EPSRC Reference: |
GR/J44018/01 |
| Title: |
A FEASIBILITY STUDY FOR VISUALLY LOCATING ROBOT GRASP POINTS ON IRREGULAR OBJECTS USING RANGE DATA |
| Principal Investigator: |
Fisher, Professor R |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Artificial Intelligence |
| Organisation: |
University of Edinburgh |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 March 1994 |
Ends: |
31 August 1995 |
Value (£): |
75,666
|
| EPSRC Research Topic Classifications: |
|
| EPSRC Industrial Sector Classifications: |
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
To demonstrate that it is possible to identify patches on an object, using several range data views, where single robot gripper fingers could make reliable and stable contacts. To develop algorithms for efficiently determining sets of contact points consistent with a pre-determined gripper kinematics.We assume that:(1) the parts are rigid and of uniform density, as that is available is visual evidence,(2) errors due to mis-representation or sensing and control systems would be corrected by an on-line robot lifting process. We do not address general robotics issues.Progress:In meeting the above objectives the following tasks have been identified:Volumetric model of the objects to evaluate kinematic constraints and avoid collision:The registered range data provides a crude volumetric model. A more sophisticated algorithm has been designed which refines the convex hull using a ray casting approach. Implementation of this algorithm is proceeding. Grasping surface modelling:We have fit plane and quadric surfaces to data using refinements such as region growing and morphological filtering. Surface normals and curvature information are extracted from this model. Surface position, curvature and normal information will be used for grasp stability evaluation. Kinematic model of test gripper:The MIT/Utah dextrous hand which has been modelled. Estimating grasp stability:To estimate grasp stability, we have adapted a wrench space approach, in which any combination of forces and moments applied to the object can be depicted. The grasp quality measure is the largest hypersphere which will fit inside the convex hull of the unit wrench. This involved the construction of a 6-dimensional convex hull (3 force and 3 torque dimensions). Grasp search algorithm:A grasping search strategy has been identified that combines a discrete search of surface patch combinations with a continuous optimisation stage once a favourable patch set has been identified. This uses the surface model and grasp quality criterion (to determine if grasps are stable), together with the kinematic and volumetric models (to determine if the grasps are reachable). This is at present work in progress. The project has 8 months to go. The planned activities in the remaining period are: (1) complete the search algorithm, (2) use and extend the volumetric model, (3) extend surface patch extraction, (4) incorporate groups work on automatic range data registration and (5) experimentation with a database of irregular, unmodelled test objects.
|
|
Key Findings |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Potential use in non-academic contexts |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Impacts |
|
| Description |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk |
| Summary |
|
| Date Materialised |
|
|
|
|
Sectors submitted by the Researcher |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
| Project URL: |
|
| Further Information: |
|
| Organisation Website: |
http://www.ed.ac.uk |