Competitive identification for self-tuning control: Robust estimation design and simulation experiments

✍ Scribed by Zdzisław Kowalczuk

• Publisher: Elsevier Science
• Year: 1992
• Tongue: English
• Weight: 686 KB
• Volume: 28
• Category: Article
• ISSN: 0005-1098
• DOI: 10.1016/0005-1098(92)90021-7

✦ Synopsis

The problem of the identification of time-varying systems by means of parallel processing in application to adaptive control is addressed. This method, previously developed for adaptive signal processing is adapted for a self-tuning control. A simple fixed (maximum-span) setting of the bank effective memory lengths is proposed. A discrete-time control-object model varying in correspondence with restricted changes in an analog prototype is considered in the simulation study. Results concerning applicability of the technique as well as its robustness to the type of system variability and to tuning parameters are shown. An improvement in the robustness to changes in the noise variance and to computational difficulties can also be obtained.