| 摘要 |
PURPOSE:To obtain an approximate value whose error is small without using many comparing means by constituting above unit so that the nearer to a value of (x) by which a curvature of a curve (y) = f (x) becomes the maximum the section is, the smaller width of the section becomes. CONSTITUTION:A value (x) - x8 obtained by subtracting a constant x8 from input data (x) by a subtracting means 100 is calculated and outputted. In this case, x8 is value of (x) by which the curvature of a curve (y) = f (x) becomes the maximum. Output data of the subtracting means 100 is data of a floating point format shown in a form of (m) X 2<n>, and (n) denotes an integer with a code of 8 bits, and at the time of -128 < (n), and the time of (n) = -128, 0.5 <= ¦m¦, and ¦m¦ < 1, respectively. In such a way, many comparing means are not required. Also, the nearer to a value of (x) by which the curvature of the curve (y) = f (x) becomes the maximum the section is, the smaller width of the section becomes, therefore, an error of output data (y) becomes small. |
