Multiresolutional recursive design operator for intelligent machines

Abstract

A general framework is proposed for the design of intelligent machines. It is based on theoretical principles that are crystallizing at the present time as a result of joint multidisciplinary efforts in the areas of robotics, intelligent control, large systems for information distribution, and decision support. This framework, called here the multiresolutional approach, reflects the nature of cognitive processes in intelligent machines. The world, the decision-making activities, the perception, and the control systems are explicitly organized within the model in such a manner as to deal with multiple representation of the same process of the intelligent machine functioning at different resolutions simultaneously. The multiresolutional approach allows for viewing all existing architectures for intelligent machines within a set of theoretical models, which makes possible their comparison and selection. It also allows for answering a number of questions that traditionally were addressed ad hoc: how many levels the hierarchy should have, how to select vocabularies at different levels, and so on.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>