Integrating Logical and Non-Logical Reasoning
Abstract
This project sets out to explore the interaction between logical and non-logical reasoning systems. It is primarily concerned with forging links between two relatively distinct sub-fields within artificial intelligence and to provide both theoretical and practical benefits. The project's fundamental goal was to explore methods for practical reasoning in the context of robotics. Robotics has seen quite significant advances in the recent past. With increases in miniaturization, computational power, sensor technology and the advent of robotic middleware (e.g., the Robot Operating System (ROS) used in this project and increasingly adopted by the major robotics research groups worldwide) we now have quite sophisticated robotic systems. However, there is a gap in higher-level reasoning and coordination of lower level behaviours. This project bridges that gap by integrating higher level logical reasoning techniques with lower level non-logical techniques. The project has made advances in dealing with erroneous information, approximate reasoning, topological reasoning and languages for practical reasoning systems. These developments are currently been implemented on a domestic robot (a Segway RMP) operating in a home environment. The project therefore demonstrates practical reasoning in complex environments.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 29, 2011
- Accession Number
- ADA549238
Entities
People
- Claude Sammut
- David Rajaratnam
- Maurice Pagnucco
- Michael Thielscher
Organizations
- University of New South Wales