• DocumentCode
    166859
  • Title

    Network MIMO with decision tree classification in downlink OFDMA networks

  • Author

    Ling Zhang ; Shenghua He ; Xiangming Wen ; Zheng Wei ; Jun Zhao ; Zhiqun Hu

  • Author_Institution
    Beijing Key Lab. of Network Syst. Archit. & Convergence, Beijing Univ. of Posts & Telecommun., Beijing, China
  • fYear
    2014
  • fDate
    4-7 May 2014
  • Firstpage
    22
  • Lastpage
    26
  • Abstract
    A coordination strategy among a large cellular MIMO network is presented in this paper. With clustered linear precoding and dynamic frequency reservation techniques, JDFR-BDGMD (Joint Dynamic Frequency Reservation and Block Diagonalization precoding with Geometric Mean Decomposition precoding scheme) greatly mitigates intra-cluster interference as well as inter-cluster interference in the downlink OFDMA networks. Furthermore, the coordination is based on decision tree classification which categorizes users into coordinative users and non-coordinative users. It is noteworthy that we pick a capacity-based subchannel from system available resources for coordinative users. Finally, an equal power allocation algorithm is utilized to distribute power. Simulation results show that our proposed scheme can significantly improve the performance of multi-cell networks.
  • Keywords
    MIMO communication; OFDM modulation; cellular radio; decision trees; precoding; JDFR-BDGMD; capacity-based subchannel; coordination strategy; decision tree classification; downlink OFDMA networks; inter-cluster interference; intra-cluster interference; joint dynamic frequency reservation and block diagonalization precoding with geometric mean decomposition precoding scheme; large cellular MIMO network; linear precoding; multicell networks; Clustering algorithms; Decision trees; Interference; MIMO; Telecommunications; Time division multiple access; Wireless communication;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Telecommunications (ICT), 2014 21st International Conference on
  • Conference_Location
    Lisbon
  • Print_ISBN
    978-1-4799-5139-0
  • Type

    conf

  • DOI
    10.1109/ICT.2014.6845073
  • Filename
    6845073