CEREBROT - Adaptive Cerebellum for Sensori-Motor Integration and its Application in Robotics
Perceiving and understanding the environment and being able to efficiently interact with it dynamically is critical for next generation of robots. The human brain is capable of processing and integrating a masive amount of hetereogeneous sensory data, both from external sources (sensory system) as from internal sources (as feedback from previously stored internal representations). Thus, the human brain is a very good reference that performs closed-loop perception-action processing tasks intensively through a highly specialized neuronal structure.
Particularly, the cerebellum plays an important role in the generation of movements in an accurate and coordinated manner (as is clear in patients suffering of any type of cerebellar ataxia). The capability of efficient multisensory integration, as well as the generation of more abstract models, generation of coordinated and accurate patterns is based on the integration of diverse plasticity mechanisms (synaptic or intrinsic mechanisms of neurons) in different layers within the cerebellum internal structure. The way these different adaptive mechanisms (distributed plasticity) complement each other is far from being understood in detail.
In this project we address the development of a system for information processing and efficient representation based on the cerebellum, and its potential application to robotics (for next generation of robotics) based on low power actuators (and inherently safe in the framework of their interaction with humans).