網路控制系統中,控制與迴授命令是透過網路作為傳輸的媒介,因而產生網路時間延遲、封包碰撞及取樣周期降低等問題。其中以時間延遲影響系統最為嚴重,此問題未作適當處理,輕則降低系統效能,重則使系統不受控制。現有的補償方法中,PID控制器擁有設計簡單的優點,但在時變系統中難有良好的控制。針對時變系統作改善,以模糊控制器取而代之,進而增強控制的適應力及避免複雜的受控模型。在時變量大的情況下,規則庫基於經驗法則來設計且為固定型,未能固定好中心點作適應型改善。所以,如何在時變量大的情況下,維持暫態響應和穩態響應是網路控制的重要課題。 因此,本論文提出一種網路控制系統之可變型時間延遲補償方法,對於時變量大的環境能維持良好的系統響應,並藉由模擬結果與實作相互驗證。其實作研究分為兩大部分,第一為即時估測時間延遲,透過監控介面調整傳送時間間隔,量測出封包往返的時間,藉此比較近端與遠端之時間延遲差異性。第二是解決時間延遲的方法,以通訊干擾估測器(CDOB)作為控制架構的基礎,探討當時變量大的情況下,通訊干擾估測器於暫態部分有些許振盪。因此控制器採用多層級模糊控制器(MLFC),依經驗法則調整第一層模糊規則庫適應機制。並以誤差與誤差變化量為系統輸入,來評斷第一層模糊控制器之系統改善有效性。再由第二層模糊控制器修改並固定中心點,來減少最大超越量及補償系統建模之不確定性,有效處理可變型時間延遲所產生的現象。
In networked control systems (NCSs), Internet serves as the transmission medium for control command and feedback command. Therefore, the transmission medium causes NCSs also lead to unavoidable problems in time delays which seriously degrade control performance and stability. Among existing time delay compensation methods, PID controller is designed to be simple, but it is difficult to have excellent control for time delay varied system. To improve time delay varied system, this project use fuzzy logic controller to replaced PID controller. Thus, such approach can enhance control adaptability and avoid complex mathematical models. When time delay varied is great, the fuzzy rules which are based on the experience of human operators and fixed type can’t be adaptive improvements. Therefore, it is critical for NCSs to be able to make transient responses and maintain steady state responses within a great time delay varied. This study proposed a NCSs design. The method can keep the system responses for time delay varied system. This study verified the result of simulation and practical experiment. The practical experiment implemented into two parts. In the first approach, on-line estimation of the time delay is developed by network monitor interface processing measurement of the round-trip time between packages. Thereby, compare of time delay difference in proximal and remote. The second approach proposed a control architecture where CDOB is the basis for time delay varied system. CDOB appears oscillations in the transient response. Therefore, multi-level fuzzy controller (MLFC) is selected as base of control response. The first layer fuzzy control rules and generated based on the experience of human operators. In order to correct the impreciseness of the first layer fuzzy control rules and compensate for system response. A second layer fuzzy controller is used to tune the output membership functions and fixed the center of control response. A systematic procedure is proposed for design of a multi-level fuzzy controller, which is capable of controlling nonlinear systems with uncertainties and time delay varied parameter to reduce overshoot and improve system performance.