著者

前花 晋作 金城 寛 上里 英輔 山本 哲彦

出版者

一般社団法人電子情報通信学会

雑誌

電子情報通信学会技術研究報告. NC, ニューロコンピューティング (ISSN:09135685)

巻号頁・発行日

vol.106, no.102, pp.85-88, 2006-06-09

---

本論文では,遺伝的アルゴリズム(Genetic Algorithm: GA)で学習するニューロ制御器(Neurocontroller: NC)を用いて,四輪車両のライントレース制御を行なう.四輪車両はdriftlessシステムと呼ばれる非ホロノミッタ系である.四輪車両のような非ホロノミック系を制御する方法として,時間軸状態制御法などのchained formへの変換を必要とする制御方法が提案されてきたが,chained formへの変換を用いる方法には,初期値に限界があるなどの問題点があった.そこで本研究では,chained formへの変換を必要としないGAで学習するNCによる制御システムの設計を行なう.

著者

松本 心 顔 玉玲 金城 寛 山本 哲彦

出版者

一般社団法人日本機械学会

雑誌

機械材料・材料加工技術講演会講演論文集

巻号頁・発行日

vol.2001, no.9, pp.355-356, 2001-11-02

---

In this paper, a diagnosis method for machine faults using a neural network based on autocorrelation coefficients of wavelet transformed signals is presented. It is important for factory engineers to accurately estimate machine faults. In conventional diagnosis methods, frequency analysis using the fast Fourier transform (FFT) has often been employed. Recently, wavelet transforms have been studied and applied to many signal-processing applications. Wavelet transforms are very useful because of characteristics of time-frequency analysis. In this paper, we propose an application of wavelet transforms to machine fault diagnosis. In order to apply wavelet transforms to machine fault diagnosis, we use autocorrelation coefficients of the wavelet transformed signal. In this research, it becomes clear that the autocorrelation coefficients, represent the different classes of machine states. For the automatic diagnosis, we trained a neural network to recognize three classes of machine states based on the autocorrelation coefficients of wavelet transformed signals. Simulation and experimental results show that the trained neural network could successfully estimate machine faults.

著者

山本 哲彦 花田 真一 中園 邦彦 金城 寛 玉城 史朗

出版者

日本機械学会

雑誌

日本機械学会論文集. C編 (ISSN:03875024)

巻号頁・発行日

vol.61, no.591, pp.4276-4281, 1995-11-25

参考文献数

3

被引用文献数

5

---

In this work we consider unstable control objects such as an inverted pendulum. Two evaluation procedures in genetic algorithm (GA) are set. The first involves the following steps : set two limits, -Θ and Θ, on both sides of the unstable equilibrium point, set an initial point θo in [-Θ, Θ], initiate a motion, measure the time when the motion reaches one limit, repeat simulations of neuro-control, select neural networks in order of length of holding times, and apply GA-crossover to superior neural networks of long holding times. The second involves the following steps : select neural networks in order of shortness of settling time to the equilibrium point, and apply GA-crossover to superior neural networks of short settling times. We adopt only the first evaluation procedure in the early generation stages of GA. After the number of neural networks of controllability reaches a sufficient percentage of all the neural networks in a computer, we adopt the second evaluation procedure, and GA evolution is continued. Neural networks of controllability appear at about the 10th generation and evolve to the ability limit predetermined by the structure of neural networks.

著者

山本 哲彦 吐合 隆拡 中園 邦彦 金城 寛 玉城 史朗

出版者

日本機械学会

雑誌

日本機械学会論文集. C編 (ISSN:03875024)

巻号頁・発行日

vol.62, no.601, pp.108-113, 1996-09-25

参考文献数

6

被引用文献数

2

---

Genetic algorithms (GAs) with rough evaluations can prompt the evolution of neural networks that are able to control unstable dynamic systems. The proposed control method exploits the advantage of GAs that time-varying evaluations can be easily incorporated. First an easy evaluation in GAs induces the appearance of neural networks with controllability. Second, an evaluation of settling time prompts the evolution of neural networks that show high performance. The method is applied to the stable control of a bicycle. Neurocontrol of the steering at direction change causes reverse response like that of a human cyclist.