Compensating high precision positioning machine tools by a self learning capable controller

Silvestri, Marco and Pedrazzoli, Paolo and Boer, Claudio and Rovere, Diego (2011) Compensating high precision positioning machine tools by a self learning capable controller. In: 11th euspen International Conference 11th euspen International Conference, 23.05.2011-26.05.2016, Como, Italy.

extended-abstract-silvestri-2.pdf - Published Version

Download (121kB) | Preview


The major part of positioning errors of multi-axis machine tools reveals systematic behaviour. Unfortunately, the dependency of these error on environmental and loading conditions passes through complex relationships for describing kinematic structure, machine elements mechanical behaviour, foundation characteristics, temperature distribution in operating conditions and workpiece mass properties. This paper describes an approach to error compensation, developed within the ongoing European research project SOMMACT (Self Optimizing Measuring MAChine Tools). This project is intended to develop and to validate an innovative production hardware and control system founded on understanding, evaluating and controlling large machine tools production performances, based on a self learning controller capable to manage a large amount of measurements coming from environmental sensors, embedded metrological systems and other useful information, all related with the referring measurements conditions. The proposed architecture uses a support vector machine kernel and a fuzzy logic inference system as the fundamental mechanisms for detecting similarities among the collected data and the current condition, because these similarities concern not comparable fields whose contribution can not a priori be added directly.

Actions (login required)

View Item View Item