Specification parameters for finite word length linear‐phase FIR single‐band PCLS filters

Summary Finite word length linear‐phase finite impulse response filters can be implemented using low‐cost devices. Among them, the single‐band peak‐constrained least squares filters are the most common one employed in many engineering applications. To design these filters, it is required to define t...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:International journal of adaptive control and signal processing Ročník 32; číslo 11; s. 1546 - 1556
Hlavní autoři: Li, Shuang, Ling, Bingo Wing‐Kuen
Médium: Journal Article
Jazyk:angličtina
Vydáno: Bognor Regis Wiley Subscription Services, Inc 01.11.2018
Témata:
Coefficients
Computer simulation
Design engineering
Filter design (mathematics)
finite word length filters
FIR filters
functional inequality constrained integer programming
linear‐phase FIR single‐band PCLS filters
Parameters
PCLS theorem
Specifications
ISSN:0890-6327, 1099-1115
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Summary Finite word length linear‐phase finite impulse response filters can be implemented using low‐cost devices. Among them, the single‐band peak‐constrained least squares filters are the most common one employed in many engineering applications. To design these filters, it is required to define the specifications on the maximum passband ripple magnitude, the maximum stopband ripple magnitude, and the number of bits for representing the filter coefficients. Most of the existing works on the filter design are on the determination of the filter coefficients based on a given set of these specification parameters. However, very few works are on the determination of these specification parameters. Hence, this paper establishes the analytical relationships among these specification parameters. In particular, the relationship between the total ripple energy and the maximum ripple magnitude of the filter, the relationship between the number of bits for representing the filter coefficients and the total ripple energy of the filter, and the relationship between the number of bits for representing the filter coefficients and the maximum ripple magnitude of the filter are established. Computer numerical simulations are presented for verifying the obtained results.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0890-6327
1099-1115
DOI:10.1002/acs.2927