For measuring the subject, the measuring points are spread on inspection feature. At this time, the number of measuring points is determined first of all. It is the most influential factor of all the inspection variables to inspection resolution. If there is a lot of measuring points, inspection time increases. Generally, for reduction of the number of measuring points and the measuring time

Fig.4 The selection of inspection mode in OMM module

(a) Manual (b) General (c) Automatic Fig.5 The distribution of measuring point each mode

Fig.6 The procedures for determination of inspection parameters

about various inspection subjects, the number of measured points are as follows : 3~5 points for a straight line, 4~9 points for a plane, 5~9 points for a sphere, 6~12 points for a cylinder and 7~12 points for a cone. Because it is the minimum number of measuring points which can define the inspection feature about subject simply, it  is the number of measuring points ruled out  the repeatability of a machine or the size of an inspection

surface. Consequently, when the geometric tolerance is appraised or the result of inspection is represented using the minimum measuring points, it is not to be trusted. In this research, the repeatability of a machine tool and the size of inspection feature are added as a factor to decide the number of measuring points. Fig.7 shows the algorithm to determine the number of measuring points using Fuzzy logic. The minimum number of measuring points which is leaded by using Fuzzy logic agrees with the number of measuring points to evaluate the geometric tolerance. It is possible to input the maximum number of measuring  points.  The  input  variables  are  the  area  of

inspection feature, the resolution to lead the result of inspection and the repeatability of the  machine tool.[9] The measuring points are altered actively by the value of input variables.

the values of s, t are larger than 0 and smaller than 1. This method is applied to the Automatic Mode. Fig.8 shows the inspection positions in case of 10 measuring points.

Table 1 Hammersley’s function for various features

252 288

Fig. 8 The distribution of ten Hammersley points on rectangular and circular surface

Fig. 7 The fuzzy logic for measuring points

3.2 Determination of the location of measuring points using Hammersley’s algorithm

The  location  of  measuring  points  to  be  applied to

OMM    module    is    determined    by    grid    type  and

3.3 Determination of the probe path using Traveling Salesperson Problem

In On-Machine Measurement, the path of probe must satisfy two conditions; generating the minimum moving distance and determining sequence of measuring

[10]

Hammersley’s  algorithm.  The  grid  type  method which

points.

The guide point minimizes the measuring error

uses the equi-interval mesh pides the range of object into definite intervals for x-, y- directions. As this mesh is applied to object, the uniform grid is made in the range of inspection object and then the intersections of x-, y- axis are decided on the inspection positions. This method is applied to the General Mode. The method using Hammersley’s algorithm establishes (s, t) axis as Hammersley’s coordinate in object. If 2D-section is a rectangle, a rectangular coordinate is used. If 2D-section is a circle, a polar coordinate is used. Table.1 shows the Hammersley’s function for various features. In this table,