Modified Shared Circuits Model for Manufacturing Processes Control: From Psychology and Neuroscience to a Computational Solution

摘要

There are many complex processes waiting for artificial cognitive solutions able to deal with new, complex, unknown, or arbitrary tasks efficiently. In this work, the modified shared circuits model (MSCM) for artificial cognitive control is presented. The main goal is to surpass the limitations of the shared circuits models and to formalize an integrated computational solution on the basis of a neuroscientific and psychological approach. Two novelties of the proposed systems are a commutation or switching mechanism between modules in order to reproduce efficiently the imitation, deliberation and mindreading characteristics of human sociocognitive skills. Another contribution is the introduction of a self-optimization strategy based on cross entropy in order to fulfil the control goals. The closed-loop behaviour of the drilling force demonstrates that the MSCM approach is an alternative and feasible option in the field of artificial cognitive control to deal with processes complexity and uncertainty.