PEM-modelling as the basis for designing intelligent robot control

Abstract

A general model for intelligent robot control is proposed. A hierarchical framework with motion and manipulation task decomposition, sensor processing and modelling sub-structures is considered as an overall system architecture. Each computational module, both related to task decomposition and sensor processing is proposed to consist of modules, for each of which there are three generic activities: planning, executing and monitoring (PEM). Planning encompasses task decomposition activities, executing activates a mechanism to carry out the plan and monitoring observes the validity of the plan related to real situations. The time dependencies of planning, executing and monitoring are described with a meta-control mechanism, for which six basic schemes are introduced. The importance within the different meta-control schemes is that they result in different overall behaviour, particularly in erroneous or unexpected situations.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>