MODELLING ROBOTIC COGNITIVE MECHANISMS BY HIERARCHICAL COOPERATIVE COEVOLUTION
Recently, many brain modelling efforts attempt to support cognitive abilities of artificial organisms. The present work introduces a computational framework to address brain modelling, emphasizing on the integrative performance of substructures. Specifically, we present an agent-based representation of brain areas, together with a hierarchical cooperative coevolutionary scheme, which is able to highlight both the speciality of brain areas and their cooperative performance. The inherent ability of coevolutionary methods to design cooperative partial structures supports the design of partial brain models and, at the same time, provides a consistent method to achieve their integration. As a result, the proposed approach proceeds in either an incremental or a compound mode. Furthermore, the performance of the model in lesion conditions is considered during the design process to enforce the reliability of the result. Implemented models are embedded in a robotic platform to support its behavioral capabilities.