SCIENTIA SINICA Informationis, Volume 51 , Issue 1 : 139(2021) https://doi.org/10.1360/SSI-2020-0006

Resource management scheme for full-duplex device-to-device vehicular communication using hypergraph clustering and interference limited area theory

More info
  • ReceivedJan 6, 2020
  • AcceptedMar 25, 2020
  • PublishedJan 5, 2021


Funded by




[1] Li Y, Jiang T, Sheng M. QoS-Aware Admission Control and Resource Allocation in Underlay Device-to-Device Spectrum-Sharing Networks. IEEE J Sel Areas Commun, 2016, 34: 2874-2886 CrossRef Google Scholar

[2] Meng L, Hua J, Lin M. Accurate joint estimation of Doppler shift and SNR in mobile communications with high vehicle speed. Sci Sin-Inf, 2017, 47: 1705-1714 CrossRef Google Scholar

[3] Ansari R I, Chrysostomou C, Hassan S A. 5G D2D Networks: Techniques, Challenges, and Future Prospects. IEEE Syst J, 2018, 12: 3970-3984 CrossRef ADS Google Scholar

[4] Xue L, Peng J, Guan X. Integrated security of cyber-physical vehicle networked systems in the age of 5G. Sci Sin-Inf, 2019, 49: 1640-1658 CrossRef Google Scholar

[5] Cheng N, Zhou H, Lei L. Performance Analysis of Vehicular Device-to-Device Underlay Communication. IEEE Trans Veh Technol, 2017, 66: 5409-5421 CrossRef Google Scholar

[6] Ren Y, Liu F, Liu Z. Power Control in D2D-Based Vehicular Communication Networks. IEEE Trans Veh Technol, 2015, 64: 5547-5562 CrossRef Google Scholar

[7] Sun W, Yuan D, Strom E G. Cluster-Based Radio Resource Management for D2D-Supported Safety-Critical V2X Communications. IEEE Trans Wireless Commun, 2016, 15: 2756-2769 CrossRef Google Scholar

[8] Sun W, Strom E G, Brannstrom F. Radio Resource Management for D2D-Based V2V Communication. IEEE Trans Veh Technol, 2016, 65: 6636-6650 CrossRef Google Scholar

[9] Sun P, Shin K G, Zhang H. Transmit Power Control for D2D-Underlaid Cellular Networks Based on Statistical Features. IEEE Trans Veh Technol, 2017, 66: 4110-4119 CrossRef Google Scholar

[10] Chen Q, Wu F, Leng S, et al. Degree of link dependence-based LTE-V2V clustering and alarm information forwarding. In: Proceedings of IEEE/CIC International Conference on Communications, 2016. 6. Google Scholar

[11] Khan A A, Abolhasan M, Ni W. An Evolutionary Game Theoretic Approach for Stable and Optimized Clustering in VANETs. IEEE Trans Veh Technol, 2018, 67: 4501-4513 CrossRef Google Scholar

[12] Zhou Z, Yu H, Xu C. Dependable Content Distribution in D2D-Based Cooperative Vehicular Networks: A Big Data-Integrated Coalition Game Approach. IEEE Trans Intell Transp Syst, 2018, 19: 953-964 CrossRef Google Scholar

[13] Xiao H, Chen Y, Zhang Q. Joint Clustering and Power Allocation for the Cross Roads Congestion Scenarios in Cooperative Vehicular Networks. IEEE Trans Intell Transp Syst, 2019, 20: 2267-2277 CrossRef Google Scholar

[14] Selvitopi O, Acer S, Aykanat C. A Recursive Hypergraph Bipartitioning Framework for Reducing Bandwidth and Latency Costs Simultaneously. IEEE Trans Parallel Distrib Syst, 2016, : 1-1 CrossRef Google Scholar

[15] Purkait P, Chin T J, Sadri A. Clustering with Hypergraphs: The Case for Large Hyperedges. IEEE Trans Pattern Anal Mach Intell, 2017, 39: 1697-1711 CrossRef Google Scholar

[16] Chen C, Wang B, Zhang R. Interference Hypergraph-Based Resource Allocation (IHG-RA) for NOMA-Integrated V2X Networks. IEEE Internet Things J, 2019, 6: 161-170 CrossRef Google Scholar

[17] Min H, Lee J, Park S. Capacity Enhancement Using an Interference Limited Area for Device-to-Device Uplink Underlaying Cellular Networks. IEEE Trans Wireless Commun, 2011, 10: 3995-4000 CrossRef Google Scholar

[18] Sun J, Zhang Z, Xiao H. Uplink Interference Coordination Management With Power Control for D2D Underlaying Cellular Networks: Modeling, Algorithms, and Analysis. IEEE Trans Veh Technol, 2018, 67: 8582-8594 CrossRef Google Scholar

[19] Liu F, Hou X, Liu Y. Capacity Improvement for Full Duplex Device-to-Device Communications Underlaying Cellular Networks. IEEE Access, 2018, 6: 68373-68383 CrossRef Google Scholar

[20] Zhang D D, Zhang Z S, Wang X. 全双工通信关键技术研究. Sci Sin-Inf, 2014, 44: 951-964 CrossRef Google Scholar

[21] Zhang G, Yang K, Liu P. Power Allocation for Full-Duplex Relaying-Based D2D Communication Underlaying Cellular Networks. IEEE Trans Veh Technol, 2015, 64: 4911-4916 CrossRef Google Scholar

[22] Xie X, Chen J, Fu Y. Energy efficiency and throughput maximization scheme based on joint antenna selection and beamforming optimization in FD-SWIPT bidirectional relay systems. Sci Sin-Inf, 2019, 49: 1217-1230 CrossRef Google Scholar

[23] Megson G M, Cadenas J. Rapid preconditioning of data for accelerating convex hull computations. Electron Lett, 2014, 48: 270-272 CrossRef ADS Google Scholar

  • Figure 1

    (Color online) Communication scenario

  • Figure 3

    Data structure

  • Figure 4

    (Color online) The diagram of clustering based on HG-C strategy


    Algorithm 1 Hyperedge coarsening and cluster dividing based on DLD at the distributed VUE


    while Beacon information do

    ${V_n}$ receives Beacon information, calculates ${D_{{V_n},{V_m}}}(t)$ by the formula of (9), creates the neighborhood list of ${V_n}$, and initializes $L({V_m},{V_n})=1$, where ${V_m}~\in~{\rm{Nei}}\left(~{{V_n}}~\right)$;

    if ${\rm{Nei}}\left(~{{V_n}}~\right)~=~\emptyset~$ then



    Calculate ${\bar~D_{{V_n}}}(t)$ by the formula of (10), and broadcast to ${\rm{Nei}}\left(~{{V_n}}~\right)$;

    Update the neighborhood list, and find out ${\bar~D_{{V_{\max~}}}}(t)$ with the max value of ADLD, ${V_{\max~}}~\in~{\rm{Nei}}\left(~{{V_n}}~\right)$;

    if ${\bar~D_{{V_n}}}(t)~>~{\bar~D_{{V_{\max~}}}}(t)$ then

    ${V_n}$ sends an join request to ${V_{\max~}}$, and resets $L({V_{\max~}},{V_n})~=~2$;


    Mark ${V_n}$ as a candidate cluster head, and form a coarsening hyperedge; then update the member list of the candidate cluster, and set $L({V_m},{V_n})~=~2$, ${V_m}~\in~c'\left(~{{V_n}}~\right)$;

    end if

    end if

    if $\forall~L({V_m},{V_n})~=~2$, ${V_m}~\in~{\rm{Nei}}\left(~{{V_n}}~\right)$ then

    $c'\left(~{{V_n}}~\right)$ is a closed hyperedge, and the vehicles in the hyperedge ${{\cal~G}_{{V_n}}}$ are divided into a cluster $c\left(~{{V_n}}~\right)$;


    Send the neighbor list and cluster member list to the base station;

    if the related cluster members receive the cluster division result of the base station then

    Send a join invitation to the corresponding member according to the member list divided by the base station, and then perform step 9;


    ${V_n}$ joins the corresponding cluster according to the invitation of the cluster head;

    end if

    end if

    end while

  • Table 1   Symbol definition
    Symbol Definition
    ${\cal~G}~=~\{~{{{\cal~G}_1},{{\cal~G}_2},~\ldots,~{{\cal~G}_g}~,\ldots~,{{\cal~G}_G}}~\}$ The number of clusters is $G$ by vehicle division in this cell
    ${{\cal~D}_g}~=~\{~{{\cal~D}_1^g,{\cal~D}_2^g,~\ldots,~{\cal~D}_d^g~,\ldots,{\cal~D}_{{N_{{C},g}}}^g}~\}$ The number of FD-D2D links is ${N_{{C},g}}$ in the $g$-th cluster
    ${\cal~D}_d^g~=~\{~{V_{\rm{h}}^g,V_{m,d}^g}~\}$ The set of VUE transceivers with $d$-th FD-D2D link in the $g$-th cluster
    ${{\cal~G}_g}~=~\left\{~{V_{\rm{h}}^g}~\right\}~\cup~\{~{V_{m,x}^g|x~=~1,2,~\ldots~{\rm{,}}{\kern~1pt}~{N_{C,g}}}~\}$ The set of VUEs in the $g$-th cluster, $V_{\rm{h}}^g$ denotes cluster-head, $V_{m,x}^g$ denotes cluster-member.
    $c\left(~{V_{\rm{h}}^g}~\right)~=~\{~{V_{m,x}^g|x~=~1,2,~\ldots~{\rm{,}}{\kern~1pt}~{N_{C,g}}}~\}$ The set consisting of $N_{C,g}$ cluster-members in the $g$-th cluster.
    ${\cal~C}_d^g{\rm{~=~}}\{~{C_{{\cal~D}_d^g}^1,C_{{\cal~D}_d^g}^2,~\ldots~,C_{{\cal~D}_d^g}^{K_d^g}}~\},{\cal~C}_d^g~\in~{\cal~C}'$ The set of the $K_d^g$ CUE spectrums of sharing for the $d$-th FD-D2D link in the $g$-th cluster.
    ${\cal~C}_g^{\rm~h}~=~\{~{{\cal~C}_1^g,{\cal~C}_2^g,~\ldots~,{\cal~C}_{{N_{D,g}}}^g}~\},{\cal~C}_g^{\rm~h}~\in~{\cal~C}'$ The set of CUE spectrums used by cluster-head multicast in the $g$-th cluster.
    ${{\cal~D}_{{C_n}}}~=~\{~{{\cal~D}_1^n,{\cal~D}_2^n,~\ldots~,{\cal~D}_{{I_n}}^n}~\},{C_n}~\in~{\cal~C}'$ The set of ${I_n}$ FD-D2D links reuses the spectrum of ${C_n}$.

    Algorithm 2 Dividing of related cluster sets based on adjacency matrix at the base station

    Require:Neighbor list and the cluster member list of candidate cluster head, $i~=~1$;

    The base station integrates the neighbor list of all candidate cluster heads to establish an adjacency matrix $A$;

    Perform block diagonalization on matrix $A$ to obtain block diagonal matrix $B$;

    Extract the non-zero blocks on the diagonal of matrix $B$ to obtain the adjacency matrix of the related cluster set ${{\cal~G}_{\rm{r}}}{\rm{~=~}}\left\{~{{{\cal~G}_{{\rm{r,1}}}},{{\cal~G}_{{\rm{r,2}}}},~\ldots~,{{\cal~G}_{{\rm{r,}}{N_{\rm~r}}}}}~\right\}$;

    while $i~\leqslant~{N_{\rm~r}}$ do

    Freely combine the column vectors of the adjacency matrix of the related cluster set ${{\cal~G}_{{\rm{r,}}i}}$, and take out the effective cluster head set ${\cal~C}_{{\rm{r,}}i}^{\rm{h}}$ with full rank;

    Find out the combination ${\cal~C}_{{\rm{r,}}i}^{{\rm{h,min}}}$ with the least number of cluster heads in ${\cal~C}_{{\rm{r,}}i}^{\rm{h}}$;

    if the number of cluster head combinations in ${\cal~C}_{{\rm{r,}}i}^{{\rm{h,min}}}$ is greater than one then

    Find out the combination with the most weighted in ${\cal~C}_{{\rm{r,}}i}^{{\rm{h,min}}}$, and delete other combinations;

    end if

    Divide VUE to the candidate cluster with the largest connection weight;


    end while

  • Table 2   Parameter
    Parameter Value Unit
    The radius of the cell 500 m
    Shadow fading 2 dB
    The SINR threshold for cellular links 12 dB
    The maximum transmit power for CUE ($P_{\max~}^{C}$) 23 dBm
    The power of Gaussian white noise $-$174mW
    Interference signal ratio at the base station (${\lambda~_B}$) 0.01
    Path loss factors (${\alpha~_0},{\alpha~_1},{\alpha~_2}$) 3, 4, 3.5
    The weighting factors for DLD (${\bf{\beta~}}$) 0.3, 0.1, 0.55, 0.05
    The number of CUEs 40, 90, 160
    The number of VUEs 500
    Maximum communication distance between vehicles 20–300 m
    Vehicle speed 40–50 km/h