The iDog project is a new umbrella project aimed at creating a semi-autonomous mobile robotic aid to help people with visual impairments. It will provide a platform for a range of subprojects, depending on students' research interests.

Tasks

The range of application tasks that could be studied include:

• Voice recognition: distinguishing an owner's voice.
• Sound localisation: what direction did a sound (or voice) come from.
• Speech recognition over a natural language subset: learning and recognising commands.
• Object recognition: recognising an owner and other important objects in a scene.
• Object tracking: for example visually following an owner.
• Text recognition: distinguishing signs from background and reading them.
• Path planning and collision avoidance: for example physically following an owner.
• Robotics: building and communicating with a motorized small vehicle.

Techniques

The projects are intended to provide a platform for developing and applying techniques from artificial intelligence (broadly interpreted to include natural language engineering and vision processing) and machine learning. Particular emphasis will be given to learning and optimisation algorithms applied to the above tasks. Examples could include:

• Evolutionary (including genetic) algorithms.
• Swarm optimisation.
• Combining global and local search.
• Neural networks.
• Statistical learning algorithms.

Students undertaking projects will become part of the Adaptive Systems Research Group. Projects will be jointly supervised with staff from other disciplines where appropriate.