Research On Device Discovery And Resource Allocation In D2D Communication

The exponential growth of smart devices such as mobile phones and tablets,the increase of the demand of online games,virtual reality(VR)and multimedia services such as high-definition video and video conferencing,pose higher challenges to the capacity,delay and security of existing networks.Device-to-Device(D2D)communication is considered to be the key technology in the future 5G mobile communication because of its proximity gain,multi-hop gain and multiplexing gain.D2 D device discovery,mode selection and resource allocation are hot topics in D2 D communication.This paper proposes a device discovery scheme based on binary backoff in Carrier Sense Multiple Access(CSMA)and two different resource allocation schemes.The device discovery scheme,based on CSMA using the binary exponential backoff algorithm,gives the device with a large number of collisions a higher priority,and the number of devices with different collision times is replaced by its expectation in every time slot.The simulation results show that the device discovery scheme can effectively reduce the average discovery delay and device discovery delay variance while ensuring high device discovery rate.The resource allocation scheme based on clustering and game theory decomposes resource allocation into two sub-problems: channel allocation and power control.Real-time small-scale fading is ignored in channel allocation and clustering is used to solve them.The power control is modeled as a non-cooperative game to maximize energy efficiency and find Nash equilibrium.The simulation results show that the D2 D users with dispersed relative distances multiply the same spectrum resources.Although the channel capacity using Nash equilibrium decreases,the energy efficiency increases.The resource allocation scheme based on greedy algorithm and Q-Learing also carries out channel allocation before power control,but each D2 D user can multiplex multiple channels.Channel allocation is solved by using greedy algorithm,power control is solved by using Q-learning.The simulation results show that when multiple channels can be multiplexed,the system capacity can be improved while maintaining high energy efficiency.