Active Vision in Autonomous Mobile Robots

This action is to support the visit of Mr Barry Bentley, from the De Montfort University, to Professor Cangelosi's Adaptive Behaviour and Cognition lab at the University of Plymouth.

Event dates: 1st Jul 2009 - 31st Dec 2009.

The exchange student will carry out new robotics experiments on active vision in autonomous mobile robots.
Active vision systems can greatly simplify the computational complexity typical, such as for navigation in unknown terrains. Past research in this field demonstrated that it is possible to use an active vision system to acquire only the relevant information from the environment in order to solve a specific task (Floreano et al., Biological Cybernetics, 2006). We will use the Mars navigation scenario as test case given the challenges to recognise and classify unpredictable features in unknown rough terrains. The new experiments will be based on the evolutionary robotics model of the Mars Rover robot (Peniak, Marocco and Cangelosi, CEC2009 Proceedings, Norway). We will extend the current sensory system model, based on local infrared sensors, to a camera-based navigation system actively controlled by the robot's own controller.