| 摘要 |
A rotation angle signal from a rotation sensor for a rotating body is converted (first conversion) into a rotation angle detection value, based on a signal value of the rotation angle signal as is, and is converted (second conversion) to a rotation angle detection value based on a signal value for where a rotation angle signal has been amplified a predetermined number of times, and according to a rotation angle region, either one of a first converted rotation angle detection value, and a second converted rotation angle detection value, is selected and output as a final rotation angle detection signal. At at least two reference rotation angles, a first learning is performed for learning a conversion characteristic in the first conversion, based on; rotation angle detection values converted by the first conversion, and reference-rotation-angle-equivalent-values set beforehand, and at at least two optional rotation angles in a rotation angle region in which the rotation angle detection value of the second conversion is selected, a second learning is performed for learning a conversion characteristic in the second conversion, based on rotation angle detection values due to conversion by the first conversion which have been learned in the first learning, and the signal value amplified a predetermined number of times.
|
