106 ニューラルネットワークを用いたゲインスケジューリング型アクティブ振動制御(GS2-2 ロボティクス・メカトロニクス)

書誌事項

タイトル別名
  • 106 Gain scheduling type active vibration control with an artificial neural network

抄録

This article examines the optimal control method of active suspension systems. Performance required in the suspension is the driving stability and ride comfort. It is well-known that there is a tradeoff between those two performance criteria. Active suspension has been attracting attention as a mechanism to cope with both ride comfort and driving stability. Even if the active control strategy is used, the tradeoff between the acceleration of the body (related to the ride comfort) and the suspension stroke (related to driving stability) still exists. In this study, we propose a new gain scheduling control method to overcome the tradeoff, that is, some fixed active controllers are scheduled depending on the sensor information to capture the state of the suspension system. The sensor information is used for the interpolation of those fixed controllers. To get the good interpolation parameter, a neural network (NN) is adopted. Various parameters in the NN are optimized with a genetic algorithm.

収録刊行物

詳細情報 詳細情報について

問題の指摘

ページトップへ