Abstract
The objective of this paper is to describe how AI (artificial intelligence) and fuzzy logic techniques can be combined to implement an on-line computer control of an axial gear-shaving process in order to improve the quality of workpiece gears. The basic principle of correcting the geometric errors of workpiece gears is introduced. The final knowledge base is obtained by the iterative procedure of trying the expert controller, observing the results and then modifying the control rules accordingly. Inference is provided by a fuzzy logic engine. Experimental results showed that accuracy of the shaved gears was greatly enhanced. The total cumulative pitch error arid the individual pitch error were reduced by approximately two times. In particular, the flank grooves, which commonly occur in the middle of tooth profiles, were removed.