Throughput Measures for Delay-constrained Communications in Fading Channels

Files in this item

Files Size Format View
Ahm2003Oct5Throughput.PDF 161.5Kb application/pdf Thumbnail

Show full item record

Item Metadata

Title: Throughput Measures for Delay-constrained Communications in Fading Channels
Author: Ahmed, Nadeem; Baraniuk, Richard G.
Type: Conference paper
Keywords: throughput; fading channels; block fading; outage; capacity
Citation: N. Ahmed and R. G. Baraniuk, "Throughput Measures for Delay-constrained Communications in Fading Channels," 2003.
Abstract: Fading channels, often seen in wireless systems, provide an unfavorable environment for reliable communications. Current methods for evaluating the performance of fading channels include ergodic capacity and epsilon-capacity. Ergodic capacity quantifies the ultimate reliable communication limit of the fading channel. It is only achievable with infinite coding delay, making it impossible to achieve in practice. epsilon-capacity, achievable with finite coding delay, does not provide a measure of error-free communications performance. Since practical communication systems are delay-constrained, it is possible to retransmit codewords when errors occur. We provide a new analysis framework that accounts for codeword retransmission in the analysis of fading channels. We introduce new measures, maximum zero-outage throughput and maximum epsilon-throughput, that predict the performance of practical systems and show that ergodic capacity and epsilon-capacity are special cases of our definitions. We also provide a measure that characterizes the performance of a system with more complex receiver design, using "incremental diversity" to improve throughput.
Date Published: 2003-10-01

This item appears in the following Collection(s)

  • ECE Publications [1048 items]
    Publications by Rice University Electrical and Computer Engineering faculty and graduate students
  • DSP Publications [508 items]
    Publications by Rice Faculty and graduate students in digital signal processing.