A Clustering-Based Approach for Low-Complexity Adaptive Profile Selection in DOCSIS 3.1

Mahdi Ben Ghorbel, Ebrahim Bedeer Mohamed, Jahangir Hossain, Colin Howlett, Brian Berscheid, Julian Cheng

Research output: Contribution to conferencePaperpeer-review

Abstract

We introduce a low complexity downlink bit-loading for each subcarrier of a large number of cable modems (CMs) in DOCSIS 3.1. Although different modulations between subcarriers are allowed in DOCSIS 3.1, the number of different bit-loading per subcarrier assignments, called profiles, is limited for computational complexity. Thus, an efficient method of determining the best profile to be used by each CM is needed. The proposed approach is based on a two-step algorithm. In the first step, users are clustered into groups based on their signal-to-noise ratio (SNR) while in the second step, the profile per group (i.e., the bit-loading per sub-carrier) is selected. Two different criteria are investigated for this step. The first one considers the average SNR among the users in the cluster to determine the bit-loading while the second one is more conservative and considers the worst SNR among the users in order to guarantee the targeted BER for all the users in the cluster. Through numerical results, we show the efficiency of the two-step approach and compare the two proposed criteria for bit-loading.
Original languageEnglish
Publication statusAccepted/In press - 30 Apr 2016
EventConference: 28th Biennial Symposium on Communications (BSC 2016) - Kelowna, Canada
Duration: 5 Jun 20167 Jun 2016

Conference

ConferenceConference: 28th Biennial Symposium on Communications (BSC 2016)
Country/TerritoryCanada
CityKelowna
Period5/06/167/06/16

Keywords

  • Adaptive Modulation
  • clustering
  • data over cable networks
  • profile optimization.

Fingerprint

Dive into the research topics of 'A Clustering-Based Approach for Low-Complexity Adaptive Profile Selection in DOCSIS 3.1'. Together they form a unique fingerprint.

Cite this