A Hybrid Expert System for Selection and Maintenance of Robotic Milling Conditions

A concept of a hybrid expert system for robot milling conditions control presented. Based on KUKA KR300 operating data. Milling conditions, measurements data, negative factors involved are considered. Expert estimations of actual milling episodes are used as criterion of successful milling. All measurable milling quality parameters are examined as possible feedback for milling conditions control. The choice is made in favor of using only real-time parameters available. Development of hybrid expert system includes method for expert estimations formalizing, system structure design, signal processing. An example of data processing and distribution within expert system given to further application in robotic milling control. The introduction of a hybrid expert control system into the robotic milling make it possible to implement two important functions for the manufacturing system. First, accumulate knowledge about the manufacturing process in form of measured data, simple integral and qualitative expert estimations - all in a solid and compact data format. Secondly, generate control signals for selection and maintenance of milling conditions using a rule's logic output system.