Robot-assisted Environmental Monitoring for Air Quality Assessment in Industrial Scenarios (RASEM)

In industrial environments, airborne by-products such as dust and (toxic) gases, constitute a major risk for the worker’s health. Major changes in automated processes in the industry lead to an increasing demand for solutions in air quality management. Thus, occupational health experts are highly interested in precise dust and gas distribution models for working environments. For practical and economic reasons, high-quality, costly measurements are often available for short time-intervals only. Therefore, current monitoring procedures are carried out sparsely, both in time and space, i.e., measurement data are collected in single day campaigns at selected locations only. Real-time knowledge of contaminant distribution inside the working environment would also provide means for better and more economic control of air impurities, e.g., the possibility to regulate the workspace’s ventilation exhaust locations, can reduce the concentration of airborne contaminants by 50%.

The RASEM project will develop a robot-assisted environmental monitoring system for air quality assessment in industrial scenarios. RASEM is enabled by continuously improving sensing and robotics capabilities. Based on sensor networks augmented by robots, drones, and potentially other mobile units (e.g., workers wearing mobile sensing nodes), RASEM will provide the capability to measure over longer times and in different places of the environment  in comparison to traditional monitoring procedures. RASEM will develop algorithms for distribution mapping of dust and gases, and novel, sophisticated exposure models in industrial environments that will enable deeper insights into long-term exposure patterns. Furthermore, these distribution and exposure models can lead to improved technical control of air impurities, ventilation systems and better safety and protection policies, and consequently, the improvement of working conditions. In this scope, RASEM supports the transformation of the industry in terms of digitalization and data analytics, with the objective to increase the safety management of complex industrial scenarios.

• Learn distribution maps of airborne pollutants in challenging industrial environments
• Optimize the combined use of low-cost and expensive sensors, and mobile and static sensors
• Represent and interpret long-term models of air quality
• Provide support for the development of mitigating counter-measures that improve air quality.

The combination of a dense sensor network equipped with cost-effective, low-fidelity sensors, with mobile robots carrying the sophisticated, high quality devices is a cost-efficient way to obtain improved environmental models. This will be achieved by learning the calibration and (dense) interpolation models between sensor nodes using measurements collected by the mobile robots.

Algorithms will be developed to create continuous 3D dust, gas, (and airflow) distribution maps, and exposure models over prolonged periods of time from distributed, heterogeneous, in-situ measurements. Visualizations of algorithm outputs will be prepared for non-experts (e.g. workers, managers, maintenance). The algorithms should include time- and event-dependency, e.g., a dependency on periodic events or a smelting process that causes a burst of dust and gas emissions. These dependencies allow extracting temporal patterns from the maps that can be correlated with changes in, e.g., the foundry operation and other seasonal changes (daily shifts, weekends), that enable the detection of abnormal situations (e.g., excessively high dust level, increased temperature) that may be used to trigger alarms.

Sensor planning with a heterogenous sensor setup: optimize the location and time of measurements combining low-cost sensors with more expensive and sparsely distributed ones, combining mobile and static sensors. 

Improve the accuracy of the estimate of human exposure to airborne chemical agents. Possibility to address challenging industrial scenarios and to mitigate health risk by means of better situation awareness.