Tracker Design for the Unknown Sampled-Data System Consisting of Randomly Switched Multi-Subsystems with Input Constraint

• Main Authors: Chi-ChunWang, 王頎鈞
• Other Authors: Jason S. H. Tsai
• Format: Others
• Language: en_US
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/88885127093376849274

碩士 === 國立成功大學 === 電機工程學系碩博士班 === 101 === In some cases, an unknown system consisting of multi-subsystems would switch from one subsystem to another passively. In this thesis, the off-line observer/Kalman filter identification (OKID) method is adopted to identify each unknown subsystem, and an active switching mechanism is presented for a passively switched multi-input multi-output (MIMO) subsystems. To resolve the input constraint problem for the unknown system, the modified observer-based model predictive control (MPC) combining with prediction-based digital redesign is proposed, without losing the good tracking performance as possible. The proposed modified model predictive control scheme can predict the future output and identify the switching instant immediately, and it systematically compresses a huge control input within the desired range by adjusting the weighting matrix of the cost function for the linear time-invariant (LTI) input-constrained system.