IoT System engineering approach using AI for managing safety products in healthcare and workplaces

Abstract

The Internet of things (IoT) has been widely used in life support facilities, especially in the healthcare sector. The management of health and safety (H&S) devices is an indispensable factor in guaranteeing human life. Thus, many companies and organisations have ensured the safety of their workplace from any overwhelming risks, such as fire, damage and accidents. The problem is that companies that have thousands of employees also have hundreds of health and safety systems in continuous use and Building Service Managers do not know which devices need maintenance or replenishment and require a significant amount of time to continuously check that their health and safety systems are maintained according to legislation requirements. The consequences of companies not meeting their health and safety obligations are serious and expensive if any of their employees are seriously injured as a consequence of the health and safety products not being in their original, immaculate order for use. Furthermore, insurance companies will refuse to pay compensation for injury to employees if it has been shown that the health and safety products are not in their original, immaculate order for use. Therefore, it makes economic sense to design and use an Internet of Things based monitoring system to record the state of health and safety products within companies. This thesis presents the design and prototype of an IoT-based health and safety (H&S) products monitoring system using artificial intelligence (AI) and image processing. The thesis aims to present a real-time control and monitoring system to determine the status of consumable hospital devices, such as first aid boxes, earplug dispensers, life jackets and fire extinguishers by using Zigbee sensors to read the measurement of the weight, and image-processing (sensor in a micro digital camera with affixed laser light) to determine the level. The proposed design is a dynamic (real-time) system that can monitor and predict the status of any equipment in hospitals or workplaces and notify the building service managers when any facilities need to be updated. Therefore, the proposed system can reduce hospitals’ and companies’ time, money and workforce requirements compared with previous manual maintenance approaches that are time-consuming and require more efforts and higher costs. Artificial intelligence infrastructure based on Ant Colony, Traveling Salesman Problem TSP and Genetic Algorithm GA is proposed to be implemented in optimisation algorithm to support salesmen in selecting the shortest path and latest time in maintaining and replenishing critically low devices at different locations in an appropriate time. Artificial neural networks are used to predict the optimal performance and the correlation between effective input factors and performance output. The achieved AI accuracy results is 96% of the proposed system which are proven to be reliable and can be used to manage many consumable H&S products in different district hospitals and companies as compared with the other approaches i.e. (Genetic, Ant-colony and TSP) that have been used at the same application whilst meeting regulatory requirements. The efficiency of the proposed system is validated because it reduces time, costs, efforts and workforce resources compared with previous manual maintenance approaches.