제어로봇시스템학회:학술대회논문집
- 제어로봇시스템학회 1995년도 Proceedings of the Korea Automation Control Conference, 10th (KACC); Seoul, Korea; 23-25 Oct. 1995
- /
- Pages.91-94
- /
- 1995
Optimal control of impact machines using neural networks
- Sasaki, Motofumi (Dep. of Mechanical System Eng., Toyama Univ.) ;
- Nakagawa, Makoto (Dep. of Mechanical System Eng., Toyama Univ.) ;
- Koizumi, Kunio (Dep. of Mechanical System Eng., Toyama Univ.)
- 발행 : 1995.10.01
초록
A newly developed discrete-time control design method for impact machines is proposed. It is composed of identification and control using neural networks, where the optimal controller with saturationn and no use of velocity measurements is obtained. By computer simulation, the proposed method is demonstrated to be effective: as the training progresses, the cost function becomes smaller, the proposed control is superior to PID control tuned with Ziegler-Nichols (Z-N) parameters; robust performance with respect to uncertainty, disturbances and working time is so good.