Research On NB-IoT Resource Scheduling Algorithm Based On Multi Service Scenarios

With the continuous development of the Internet of Things(IoT)technology,NB-IoT related industries have sprung up,and major mobile operators have accelerated the construction of NB-IoT base stations.NB-IoT technology has become the preferred long-distance wireless communication IoT technology for various industries due to its large capacity and low power consumption.Since the effective bandwidth of NB-IoT has limited to 180kHz,the lacking wireless resource turns into a bottleneck for NB-IoT in the ultra-large connection performance.Under the circumstance of unexpanded wireless resource,traditional resource scheduling algorithms can neither meet the diversified QoS requirements generated by the industry's horizontal expansion nor satisfy the number of ultra-large connected devices in the 5G era.Therefore,optimizing resource allocation has become an important approach to further increase the number of system connections.This article analyzes the current prevailing demands of NB-IoT,starting with accessing large-scale access scenarios simultaneously and dynamic access scenarios;improving resource scheduling algorithm tailored to the varied needs on the basis of the original algorithm.The main work is as follows:Firstly,this paper studies the shortcomings of traditional resource scheduling algorithms for NB-IoT in various business scenarios and summarizes the existing QoS demands.Moreover,it outlines the existing access restriction algorithms in dynamic access scenarios.Secondly,under the circumstance of accessing large-scale access scenario simultaneously,this paper proposes corresponding system communication models for abnormal reporting services and periodic reporting services.It solves the problem with the goal of maximizing the number of connections of a single base station at the same time.For abnormal reporting services,in order to satisfy QoS requirements,devices are divided into delay-sensitive devices and non-delay-sensitive devices.For delay-sensitive devices,considering the maximum transmission delay and the needed amount of data packets to upload,the paper proposes an improved maximum carrier-to-interference ratio algorithm.For non-delay-sensitive devices,combined with the idea of clustering in machine learning,the paper put forward a joint clustering algorithm and the reuse of equipment resources.For periodic reporting services,this article adds the attribute of the number of retransmissions to each device and proposes a priority algorithm based on the number of retransmissions.The simulation proves the advantages of the proposed algorithm compared to the existing resource scheduling algorithms.The results show that the improved algorithm can increase the number of successfully served devices while satisfying QoS requirements.Finally,the paper proposes a dynamic access restriction algorithm based on Q-learning by establishing a dynamic access traffic model of the NB-IoT in the dynamic access scenario and combining the advantages of reinforcement learning and continuous interactive learning of the environment.The Q table obtained through training dynamically adjust the access restriction factor value under different congestion conditions.Simulation results indicate that the proposed algorithm could improve system performance and reduce device access delay.