Spectrum Resource Allocation And Access Point Selection For LiFi Network

Visible Light Communication(VLC)has become a promising access technology in the field of indoor wireless communication because of the advantages of huge unregulated bandwidth resource,low power consumption and environmental friendly.LiFi(Light-Fidelity)using LED lighting as access point(AP)forming a wireless network of very small optical attocells has become a recent research hotspot.In order to solve the problems of co-channel interference and communication interruption in LiFi network,this thesis focuses on the allocation of spectrum in LiFi network and AP selection in LiFi/WiFi heterogeneous network.The downlink performance of LiFi network is analyzed.A system level downlink framework based on DCO-OFDM is introduced.The Monte Carlo simulation is applied to get the SINR statistics with HEX and PPP cell deployments.The performance of system in terms of data rate is evaluated.A one-dimensional extended model of LiFi network in special scenes is constructed.The effect of frequency reuse factor on the performance of SINR and data rate is discussed.For the allocation of spectrum resources in LiFi network,a fractional frequency(FFR)reuse technique in LiFi network is discussed.Taking the practical application scene as the research object,rectangle cell deployment model of optical communication applied to subway carriage is proposed.Two FFR schemes,strict fractional frequency reuse and soft frequency reuse,are considered in the model.The performance of SINR statistics and the spectral efficiency of the optical cellular system are analyzed.A recommended scheme is given for constructing the visible light communication network in subway carriage.For the allocation of access point in indoor LiFi/WiFi hybrid network,the dynamic change of the number and location of users in the network is fully considered.Upon the foundation of dynamic load balancing algorithm based on data rate threshold,two kinds of AP allocation methods,dynamic handover threshold and minimum data rate constraint,are proposed.The performance of outage probability and data rate of the system are analyzed.In the case that the future network state of indoor LiFi/WiFi hybrid network is dynamic and unknown,based on Q-learning in reinforcement learning algorithm,access point selection strategy is obtained by means of continuous interactive real-time learning between target user and the environment.A network model based on SINR distribution is constructed.Considering that mobile users may experience frequent handovers in optical attocell network,a reasonable setting of handover loss factor can ensure the performance of user data rate and reduce handover times.Introducing case-based reasoning algorithm,the convergence speed of Q-learning algorithm is accelerated.