Cross-layer optimization for multimedia transport over multicode CDMA networks

Hui Chen, Chun Bun Henry Chan, Victor Leung

Research output: Journal article publicationJournal articleAcademic researchpeer-review

5 Citations (Scopus)


Most previous work on code division multiple access (CDMA) considers the bit error ratio (BER) at the physical layer and the frame drop ratio (FDR) at the medium access control layer separately. However, a better system performance, e.g., in terms of a lower overall frame loss ratio (FLR), can be achieved if BER and FDR are jointly optimized. In this paper, we propose a cross-layer optimization scheme called traffic-adaptive scheme for multicode CDMA operating over a time division multiple access (TDMA) channel. Based on the traffic condition and buffer status, this scheme employs a Markov Decision Process (MDP) to determine the optimal value of the maximum number of simultaneous data frames that can be transmitted in each time slot of a TDMA frame so as to minimize the overall FLR of the system. To facilitate implementation, we also propose an approximation scheme named the rate-adaptive scheme to reduce the computation cost. Simulation and analytical results show that both the traffic-adaptive scheme and rate-adaptive scheme can significantly reduce FLR, increase the system throughput, and optimize the packet access delay of the system. Furthermore, the rate-adaptive scheme can achieve a performance close to the traffic-adaptive scheme when the traffic load in the system is high.
Original languageEnglish
Article number5669316
Pages (from-to)810-820
Number of pages11
JournalIEEE Transactions on Mobile Computing
Issue number6
Publication statusPublished - 1 Jun 2011


  • multicode code division multiple access
  • Multimedia communication
  • wireless

ASJC Scopus subject areas

  • Software
  • Computer Networks and Communications
  • Electrical and Electronic Engineering


Dive into the research topics of 'Cross-layer optimization for multimedia transport over multicode CDMA networks'. Together they form a unique fingerprint.

Cite this