Learning Control

We specialise in a number of applications involving learning of perception-based control in autonomous systems such as robots. This includes fast learning of vision-guided behaviours for mobile robots using reinforcement learning, where the learning is based on trial and error through interaction with the environment, and artificial evolution of flight controllers for aerial robots. Other work involves biologically-inspired behaviour learning, for example, using models of locust visual neurons for collision avoidance in autonomous vehicles.
Investigators

• Marwan Shaker
• Mark Smith
• Shigang Yue
• Tom Duckett

External Cooperation

Dept. of Physics, Systems Engineering and Signal Theory, University of Alicante,

Representative Publications

Shigang Yue, Reactive direction control for a mobile robot: A locust-like control of escape direction emerges when a bilateral pair of model locust visual neurons are integrated , to appear in Autonomous Robots

Marwan Shaker, Tom Duckett and Shigang Yue, Vision-Based Docking using Approximate Policy Iteration, 14th International Conference on Advanced Robotics (ICAR 2009), Munich, Germany, June 22th-26th, 2009.
 

Martínez-Marín, T. and Duckett, T. (2008): Learning Visual Docking for Non-Holonomic Autonomous Vehicles, Proc. IEEE Symposium on Intelligent Vehicles (IV08). IEEE Intelligent Transport Systems Society, Eindhoven, Netherlands, June 4-6.

Yue, S. and Rind, F.C. (2006): Collision detection in complex dynamic scenes using a LGMD based visual neural network with feature enhancement, IEEE Transactions on Neural Networks, Vol. 17, No. 3, pp.705-716.