AnExperimental StudyofNetwork-Based DC MotorSpeed Control using SANFIS YodyiumTipsuwan yytgku.ac.th
Jirat Srisabye · 2007
Network-based control (NBC)systems can provide several advantages amongtraditional control systems. Nevertheless, theperformances ofNBC systems canbedegraded duetoundesired networkbehaviors suchasnetwork-induced delays. Several NBC algorithms usually neglect several networkbehaviors duetoas- sumptions inproblem formulations. Theincompleteness andambiguity ofthisnetworkinformation implies am- biguities inNBC performances. Inthis paper, weapplied a novelNBC gainscheduling schemebyapplying a SANFIS(Self-Adaptive Neuro-Fuzzy Inference System) alongwithgainscheduling tohandle ambiguities innet- workbehaviors. TheSANFISisutilized toclassify acur- rentnetwork condition inordertoselect anoptimal gain forthis condition. Anexperimental result showsthatthe PIcontroller withtheproposed approach yields signifi- cantly better NBC performances. Keywords-Network, control, delay, neural network, adaptive system