Research On Adaptive Method In Industrial Internet Based On Machine Learning Algorithms

With the rapid development of the new generation of information and communication technology,wireless communications and artificial intelligence have entered a new research phase.It has brought a revolutionary impact on the global industry.Meanwhile,the concept of Industrial Internet has emerged.With the strong support for relevant research at the national level such as German Industry 4.0,American Industrial Internet,and China's “Made in China 2025”,the Industrial Internet and related fields have received extensive attention and the development momentum is very active.In the process of data collection and transmission in the Industrial Internet,with massive industrial devices connected to the network,the packets may generate a large amount of congestion,causing network performance problems such as transmission delay and increase in packet loss rate,affecting the industrial data real-time transmission.At the same time,due to the problems in industrial environment,network architecture,and resource allocation management models,the system energy consumption is high and the network resource utilization is low.In order to solve the above problems,this paper first focuses on the joint management of network and computing resources in industrial wireless networks.The spectrum resource allocation problem is modeled as a dynamic stochastic optimization problem.Then using POMDP to describe the dynamic parameters and solve the optimal spectrum allocation action.Afterwards,this paper focuses on the problems of real-time data transmission and network energy consumption in industrial networks,and models the multi-objective optimization problem of transmission delay and network energy consumption as the gateway optimal location process.Finally,based on the characteristics of industrial sensor networks and KMeans clustering algorithm,the KMSA algorithm is proposed to solve the optimal number of gateways and deployment location.Based on the key technologies of the LPWAN and machine learning algorithms,this paper conducts an in-depth study of the problems in the process of data transmission in the Industrial Internet and proposes adaptive algorithms.First,the joint resource management scheme is proposed to solve the problem of data transmission and processing in the network layer.The second part proposes a multi-objective optimization solution for data transmission and energy consumption in the sensing layer of the industrial wireless network.And the main research works are described as follow:(1)Joint resource management in cognitive radio and edge computing based industrial wireless networksAiming at the shortage of spectrum resources and the demand for real-time transmission and processing of industrial data in existing industrial networks,an industrial wireless network architecture based on cognitive radio and edge computing is proposed.The routers in the networking layer have both cognitive radio and edge computing.The cognitive radio technology can help the industrial network to collect the spectrum resources and expand the available frequency band to solve the problem of shortage of spectrum resources.Meanwhile,edge computing is used to solve the data processing request of the packet in the transmission process,so as to alleviate the computational pressure of the cloud server.In addition,a joint management scheme for network and computing resources is proposed and modeled for network resource allocation actions.Finally,the POMDP is used to solve the problem of dynamic random parameters in the model.The simulation results show that the scheme proposed in this paper can significantly improve network performance,solve network congestion problems and reduce the transmission delay of industrial data.(2)An adaptive method based on K-Means clustering algorithm in LPWANAiming at the real-time transmission requirements of delay-sensitive data in industrial wireless networks and the optimization of network energy consumption,a three-layer wireless network architecture based on LPWAN is proposed.The intelligent gateways deployed in the transmission layer can use edge computing to assist the cloud server in preprocessing the data on the industrial site.Meanwhile,sensor nodes with adaptive data rate technology in the sensing layer can adaptively adjust the transmission rate and power according to the transmission distance.In addition,a multi-objective optimization scheme of network energy consumption and transmission delay is proposed,and converted into a dynamic optimal location selection process of the intelligent gateways.The KMSA algorithm is proposed based on the characteristics of industrial sensor networks and the K-Means clustering algorithm.Using the KMSA algorithm to optimize the number of gateways and deployment locations.The simulation results show that the proposed scheme is obviously better than the existing ones.It can optimize the network energy consumption under the premise of guaranteeing the realtime transmission of delay-sensitive data,improve the balance of network load,and prolong the network life cycle.