Truncation of fractional derivative for online system identification
Language
EN
Article de revue
This item was published in
IFAC-PapersOnLine. 2022-01-01, vol. 55, n° 34, p. 60-65
English Abstract
Fractional derivatives are non local operators that has compacity property in terms of parameter number for modeling diffusive phenomenon with very few parameters. One of its main properties is its non-local behavior, as ...Read more >
Fractional derivatives are non local operators that has compacity property in terms of parameter number for modeling diffusive phenomenon with very few parameters. One of its main properties is its non-local behavior, as it can be exploited to model long-memory phenomena such as heat transfers. However, such non-locality implies a constant knowledge of the full past of the function to be differentiated. In the context of real-time system identification, this may limit the experiences as calculations become slower as time progresses. This study deals with the relationship between frequency content of a signal and its truncation error in order to obtain real-time exploitable algorithms.Read less <
Keywords
Fractional derivative
Short memory principle
LMRPEM
System identification