Study On Distributed Queuing And System Performance For LoRaWAN

ALOHA mechanism is adopted to manage shared channel access in the existing LoRaWAN system.However,as the inflow load increases,congestion and collision caused by data transmission occurs,leading to the decrease of throughput performance and energy efficiency.Compared with ALOHA and CSMA,time-division based distributed queuing(DQ)scheduling can not only eliminate access collisions,but also achieve stable transmission performance and scalability.Based on the analysis of channel access in LoRaWAN network,this thesis studies a new scheme of MAC layer combining with DQ algorithm,and analyses the throughput,delay and energy consumption of the system.Firstly,the thesis comprehensively discusses the random conflict resolution in wireless channels,LoRa technological foundation and the related principle of DQ.The LoRaWAN model is established to simulate the functions of the network,and the simulation program is designed based on NS3 to study how the nodes communicate in the network and analyze the overall transmission performance of the system.Secondly,aiming at the problems of data collision and channel contention when communicating in LoRa,a new scheduling scheme for channel conflict resolution is designed based on DQ.A multi-step channel access control flow is designed from the perspective of the gateway and end-device.Meanwhile,for the two scenarios of successful access and concurrent conflict,conflict resolution mechanism and the theoretical calculation methods of throughput performance are analyzed.Finally,considering the actual multi-terminal access scenario,the performance of the proposed channel contention solution is evaluated.The throughput is reckoned theoretically;the communication model is established to analyze the delay and the energy efficiency.The results through numerical calculations show that when the number of contention slots is not less than 3,the throughput can reach 74%,which is far above ALOHA;the latency can be reduced by 38% compared with M/M/1 queue;and the energy consumption can decrease up to 40% compared with the reported scheme when 10000 terminals are accessed simultaneously.