Research On Traffic Aware Transmission Resource Allocation And Edge Computing Optimization Technology In Internet Of Things

The Internet of Thing(IoT)has become an important part of the new generation of mobile communication networks,but with the increase in the size of IoT devices,coexistence management problems between massive machine type communication(mMTC)and traditional human type communication(HTC)and the large amount of data generated by IoT devices have increased the pressure on communication network traffic.The current computing scheme is hard to meet the requirements of real-time requirements for data processing results in typical mMTC scenarios such as smart factories.This paper studies the communication problem in the coexistence interference scenario of mMTC,and designs a new edge computing offloading scheme based on convolutional neural networks(CNN).The asymmetry traffic between downlink(DL)and uplink(UL)in mMTC systems is so prominent that it makes the traditionally fixed frame protocols insufficient.Meanwhile,the dynamic time-division duplexing(D-TDD)is a promising and attractive technology since its amount of time slots for the DL and UL can be asymmetric and adjusted dynamically.In the cellular and mMTC co-existing network,to balance the discrepancy between UL and DL and alleviate the interference as well,this study designs a D-TDD based transmission frame structure to first fulfill the basic transmission requirements of the HTC users with a low power almost blank subframes(LP-ABS).Herein,stochastic geometry methods are adopted to calculate spectral efficiencies of the HTC user equipment(UE).Then,focusing on the worst cluster head(CH)queue state in mMTC,we devise the slot allocation problem with the min-max objective of UL/DL queues and utilize the sub-gradient descent(SGD)method for solution.Simulation results show that the proposed traffic aware subframe configuration is more appropriate for the dynamical asymmetry environment.Meanwhile,the adopted dynamic step sized SGD algorithm can achieve a tradeoff between worst-case queue and the network throughput.In addition,this paper takes the typical application example of mMTC-Intelligent Factory as the computing scenario,uses the CNN model to process image data,and adopts the special hierarchical structure of deep neural network to combine CNN with edge computing technology,and proposes a new layered offloading scheme.In this paper,the quantitative relationship between the computational quantity and the output data by each layer of the neural network is fully considered,and the upper layer calculation of the neural network is offloaded to the cloud for continuous execution.In order to reduce the pressure of network traffic and make full use of computing resources to improve the timeliness of data processing,this paper builds an optimization problem with the computing resource allocation of the offloading scheme based on the queue model and the power consumption model,and solves the optimization problem using the reinforcement learning method-deep deterministic policy gradient(DDPG)algorithm.The simulation results show that this solution greatly reduces network throughput,relieves network transmission pressure,and achieves a good balance between data transmission and computational power consumption.The numerical simulation results show that the proposed new computation offloading scheme based on CNN computing model achieves excellent performance in energy saving and average computation delay in comparison with conventional edge computing and cloud computing scheme.