Transmission Lengths That Maximize Throughput of Variable-Length Coding & ACK/NACK Feedback
Conference Paper
-
- Overview
-
- Identity
-
- Additional Document Info
-
- View All
-
Overview
abstract
-
© 2018 IEEE. Variable-length (VL) coding sends an initial codeword followed by subsequent transmissions of incremental redundancy (IR) sent when the decoder indicates through feedback that it has not yet identified a reliable codeword. VL coding is a staple of modern communication to handle fading, and recent theoretical analysis and applications have demonstrated its value on non-fading channels for applications that require short blocklengths. To maximize throughput in a VL setting, the length of each IR transmission should be optimized. Sequential differential optimization (SDO) computes transmission lengths that optimize throughput by minimizing average blocklength. SDO produces a family of solutions that each maximize throughput for a specified maximum number of transmissions. This paper considers the average number of feedback transmissions per message as an alternative metric for the cost of the feedback resource. A Lagrangian approach provides a new SDO solution that jointly minimizes both the average blocklength and the average number of feedback transmissions associated with a message. The mapping of real-valued SDO solutions to the necessarily integer transmission lengths is also addressed.
name of conference
-
GLOBECOM 2018 - 2018 IEEE Global Communications Conference
published proceedings
-
2018 IEEE Global Communications Conference (GLOBECOM)
author list (cited authors)
-
Wesel, R. D., Wong, N., Baldauf, A., Belhouchat, A., Heidarzadeh, A., & Chamberland, J.
citation count
complete list of authors
-
Wesel, RD||Wong, N||Baldauf, A||Belhouchat, A||Heidarzadeh, A||Chamberland, J-F
publication date
publisher
published in
Identity
Digital Object Identifier (DOI)
International Standard Book Number (ISBN) 13
Additional Document Info
start page
end page
volume