| EPSRC Reference: |
EP/C530683/1 |
| Title: |
Abductive Robot Perception: Modelling Granularity and Attention In Euclidean Representational Space |
| Principal Investigator: |
Shanahan, Professor M |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Electrical and Electronic Engineering |
| Organisation: |
Imperial College London |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 December 2004 |
Ends: |
30 November 2006 |
Value (£): |
325,715
|
| EPSRC Research Topic Classifications: |
| Artificial Intelligence |
Image & Vision Computing |
|
| EPSRC Industrial Sector Classifications: |
| No relevance to Underpinning Sectors |
|
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
The purpose of the proposed project is to investigate embodied perception and reasoning in the context of Cognitive Robotics, the application of the principles and practice of knowledge representation, formal logic and reasoning to embodied, physical, robotics. We propose to investigate these topics in the immediate context of Shanahan's Theory of Abductive Perception ([Shanahan, 2002; 2004b; Shanahan and Randell, 2004]), which models perception as an abductive reasoning (reasoning from observations to possible causes) task. The Theory of Abductive Perception presents a wellfounded logical treatment for contemporary issues in machine perception.We identify three essential areas for further detailed investigation, leading to advances in the principles and practice of Cognitive Robotics. First, granularity, providing the basis for working with scalable, computationally tractable and task relevant logical models while also providing the ability to reason with descriptions of the world at varying levels of detail, matching the limited resolution of robotic sensors. Second, attention, the bi-directional mechanism by which the robot's reasoning processes are focused on salient features in its environment and by which the perceptual process is itself driven by the robot's tasks and expectations. Third, we propose to support these advances with a flexible approach to the representation of the threedimensional space a robot must both operate in and reason about, based on the notion of point clouds of abducible sensor derived features. These encapsulate the balance between a detailed data rich low-level representation for detailed prediction with information rich abstractions for high-level reasoning, and provide a clear path between the two.
|
|
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.imperial.ac.uk |