Research On Distributed Real-time Scheduling In Industrial Wireless Network

Most traditional manufacturing factories use wired networks for network deployment.Although this deployment method is mature in technology,its scalability is poor and the entire network system is difficult to maintain,making it difficult to adapt to the new industrial manufacturing environment.Compared with traditional wired networks,industrial wireless networks only need to deploy a small number of cables to deploy a certain number of wireless nodes in an industrial environment for sending and receiving data,and deployment becomes more flexible.For wireless industrial networks,the International Organization for Standardization has formulated a series of communication standards,which provide strong support and promotion for the actual deployment and use in factories.In response to the problems encountered in the actual environment,the academic community has proposed a series of optimization design methods for industrial wireless networks.Among them,the optimization design method of media access control based on scheduling is the most popular research direction at present.Based on the distributed scheduling optimization method,this paper proposes a distributed industrial wireless network real-time scheduling strategy,and conducts a series of related studies.The main contents are summarized as follows:(1)This paper proposes an asynchronous deterministic wireless communication scheduling model under the condition of time slot boundary alignment.Using this model,communication can be completed with any time slot offset between different nodes,so there is no need to synchronize clocks between nodes.Then,a method is proposed to convert the time slot aligned model into the time slot unaligned model,so that the asynchronous deterministic wireless communication scheduling model can be applied to any environment.(2)Due to the probabilistic packet loss in the process of wireless communication,this paper further proposes an asynchronous communication scheduling model with probabilistic characteristics,which can meet the reliability requirements of industrial wireless networks.After analyzing the above-mentioned asynchronous communication scheduling model,we found that the scheduling scheme required by the generation model is an NP-complete problem,and the required scheduling scheme cannot be obtained using conventional methods.Therefore,we combined the classic artificial intelligence algorithm-Monte Carlo tree search algorithm,and proposed an industrial wireless network scheduling pattern generation algorithm based on the improved Monte Carlo tree search algorithm.However,because the classic Monte Carlo tree search algorithm has a large search space and computational complexity,we combine random sampling and bit operations with the improved Monte Carlo tree search algorithm to improve the reliability of the calculation results.The calculation speed is reduced,and the algorithm running time is reduced.(3)The proposed algorithm for industrial wireless network scheduling pattern generation based on the improved Monte Carlo tree search algorithm is compared with the classic Monte Carlo tree search algorithm.The proposed algorithm calculates the average value,maximum value,and minimum value of the results in each iteration.Both value and standard are better than the classic Monte Carlo tree search algorithm.The average result of the improved Monte Carlo tree search algorithm is about 5%higher than that of the classic algorithm,the minimum is about 8%,and the standard deviation is about 40 times higher.This paper uses the simulation software cooja of the Contiki platform to conduct simulation communication experiments on the scheduling scheme generated by the improved Monte Carlo tree search algorithm.The communication reliability rate obtained by the experiment is basically consistent with the theoretical calculation value,which proves the application prospect of the asynchronous communication scheduling model.