A sensor scheduling method for industrial network systems based on edge computing

Abstract

The invention discloses a sensor scheduling method of an industrial network system based on edge computing, and relates to the field of industrial network systems. The method includes the following steps: setting up N edge computing units between n sensors on the device layer and the cloud, and connecting each edge computing unit to all sensors; if there is an edge computing unit whose number of sensors does not meet the connection number requirements, then Propose a new scheme, calculate the total cost of the new scheme, and select the scheme with the smallest total cost as the current scheduling scheme until the number of sensors connected to each edge computing unit meets the connection data requirements. At this time, the final sensor scheduling scheme for all edge computing units is obtained. The present invention considers the impact of estimation accuracy, sensor activation energy consumption, calculation delay, etc., effectively reduces the calculation pressure of the industrial cloud, has the advantages of low delay, privacy and security, and can obtain a more reasonable sensor scheduling scheme than traditional methods.

Description

technical field [0001] The invention relates to the field of industrial network systems, in particular to an edge computing-based sensor scheduling method for an industrial network system. Background technique [0002] The industrial Internet of Things collects data such as production equipment and products to the cloud platform for calculation, and then uses machine learning and other methods for analysis and prediction. With the proliferation of access network devices, a large amount of data transmission is required, which puts a lot of pressure on industrial network systems. Edge computing can provide a miniature data processing center for industrial networks, process local data, and then transmit it to the cloud, thereby improving the efficiency of connecting IoT devices to the cloud and reducing the pressure on the cloud. In addition, because edge computing is closer to devices and users, it also has the advantages of low latency, high reliability, and privacy security...

AI Technical Summary

Problems solved by technology

The domestic application number is CN111459627A, and the patent titled "Intelligent Terminal Sensor Scheduling Method Based on Reinforcement Learning" uses a sensor scheduling method based on reinforcement learning. Although it can reduce sensor energy consumption while ensuring accuracy, it still uses a single Layer sensor network, without edge computing layer, will bring a lot of pressure to the cloud in the face of massive equipment and data in the industrial background

The domestic application number is CN112416532A and the patent titled "A System and Method for Industrial Data Processing" adopts the idea of ​​edge computing, and mainly involves how the edge computing platform processes data from terminals, but it does not involve how multiple terminals can choose How to choose the terminal scheduling method in the scenario

Images