A Simplified Cerebellum-Based Model for Motor Control in Brain Based Devices

Abstract

The cerebellar system is implicated in motor learning for movement coordination. In this paper, we suggest a simplified cerebellar model with priority-based delayed eligibility trace learning rule (S-CDE) that enables a mobile agent to randomly navigate in an environment. The depth information from a simulated laser sensor is encoded as neuronal region activity for velocity and turn rate control. A priority-based delayed eligibility trace learning rule is proposed to maximize the usage of input signals for learning in synapses on Purkinje cell and cells in the deep cerebellar nuclei. Asymmetric weighted sum and velocity signal conversion algorithms are designed to facilitate training in an environment containing turns of varying curvatures. S-CDE is developed as a brain-based device and tested on a simulated mobile agent which had to randomly navigate maps of Singapore and Hong Kong expressways.