This chapter discusses the design of unknown input observers for a class of time-delay systems with a constant time delay. Based on our newly proposed matrix decomposition technique, we can effectively obtain a reduced-order disturbance-free time-delay system, and then obtain a closed form analytical expression for the unknown input. Here, the unknown input is in the simplest form containing only one output measurement, its delayed output measurement, the first-derivative of one output measurement, and a generalized functional that contains a state-delayed term, i.e., \(z(t)=Fx(t)+F_dx(t-\tau )\) . Thus, the unknown input can be estimated if we can estimate this generalized functional. We extensively look at possible ways to realize the estimation of the generalized functional by examining various design options and scenarios. We then conclude that an effective and economical way is to estimate the generalized functional directly.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Unknown Input Observers Design for Time-Delay Systems

  • Hieu Trinh,
  • Van Thanh Huynh,
  • Samson Yu,
  • Tyrone Fernando

摘要

This chapter discusses the design of unknown input observers for a class of time-delay systems with a constant time delay. Based on our newly proposed matrix decomposition technique, we can effectively obtain a reduced-order disturbance-free time-delay system, and then obtain a closed form analytical expression for the unknown input. Here, the unknown input is in the simplest form containing only one output measurement, its delayed output measurement, the first-derivative of one output measurement, and a generalized functional that contains a state-delayed term, i.e., \(z(t)=Fx(t)+F_dx(t-\tau )\) . Thus, the unknown input can be estimated if we can estimate this generalized functional. We extensively look at possible ways to realize the estimation of the generalized functional by examining various design options and scenarios. We then conclude that an effective and economical way is to estimate the generalized functional directly.