Biologically inspired computation for real-time motor control
Paul Chadderton (Bioengineering)
Claudia Clopath (Bioengineering)
An important area of robotics research is focused on developing machines that can learn and adapt to changing environments as efficiently as animals. Biological principles underlying the acquisition of motor skills are likely to inspire these new technologies. The aim of our project is to characterise computational principles of motor learning in real time using novel neurophysiological data from a uniquely tractable and well-controlled system, the whisker circuitry of the cerebellar cortex. Neural recordings will reveal how the brain adapts during learning, and will then be used to develop and apply a biologically constrained computational model in an actively sensing robot.