He Qijun
There is dust pollution in industrial production workshops, which is one of the main causes of occupational lung diseases. Traditionally, the methods used to measure dust mostly focused on a specific area or only measured the concentration occasionally, failing to present the actual exposure situation of workers in complex environments. This document presents an optical sensor network system with multiple channels, combined with worker time and movement path analysis, and then combined with graph neural network technology to construct the corresponding assessment and expectation framework. The first step is to use distributed optical sensors to form a spatial field of dust concentration in the workshop, thereby recording the dispersion of particles and having a very high resolution; the second step is to combine indoor positioning technology to reproduce the path that each worker has walked, and then estimate the cumulative damage to their entire respiratory tract in that concentration field, thus completing the calculation process from both time and space aspects. Remove the situations such as when workers encounter each other, changes in their positions, and the amount of things they have come into contact with as data in the graph structure, and then use the characteristics of graph neural networks that are good at handling relationships to create a dynamic alert model for occupational disease hazards. This model not only considers how much exposure a single person has received, but also finds the hidden patterns of the spread and accumulation of risks within the population, providing new methods for the comprehensive improvement of the occupational health protection.
Occupational Dust Exposure, Multi-Channel Optical Sensing, Spatio-Temporal Assessment, Graph Neural Networks (GNN), Dynamic Risk Early-Warning