@article{oai:air.repo.nii.ac.jp:00001524,
 author = {三浦, 武 and 秋山, 宜万 and 谷口, 敏幸 and MIURA, Takeshi and AKIYAMA, Yoshikazu and TANIGUCHI, Toshiyuki},
 journal = {秋田大学工学資源学部研究報告, Scientific and technical reports of Faculty of Engineering and Resource Science,Akita University},
 month = {Oct},
 note = {In the drive of a stepping motor, it is well known that rotor oscillation at the natural frequency causes harmful phenomena like low-frequency resonance, acoustic noise, etc. Electrical damping techniques in which the switching sequence of each phase is changed in respective manners are often used for avoidance of such problems. In these methods, the exciting period of each phase must be tuned appropriately, and several adaptive methods for on-line tuning of the exciting period have been suggested.
The authors suggest the method for estimation of rotor angle from the data of exciting currents in stator windings using a feedforward neural network, with the regulator type on-line tuning method for half-step type damping sequence. In this method, direct detection of the rotor angle can be omitted, and the advantage of non-sensor drive of a stepping motor is maintained. It is demonstrated experimentally that the exciting period is tuned appropriately and rotor oscillation is suppressed by this method under the driving conditions with different inertial loads which causes the variation of oscillatory characteristic.},
 pages = {43--51},
 title = {ニューラルネットワーク角度推定器を用いたステッピングモータ励磁タイミングの調整},
 volume = {21},
 year = {2000}
}