Predicting the Risk of Electromagnetic Interference in Hospital Equipment Using Fuzzy Neural Networks
Publication Type
Conference Proceeding
Publication Date (Issue Year)
2023
Journal Name
2023 Photonics and Electromagnetics Research Symposium, PIERS
Abstract
Healthcare facilities are populated with electronic medical devices and many other wireless electronic devices for work purposes and for personal use. As they operate, electronic devices emit electromagnetic signals raising concerns of possible electromagnetic interference (EMI) to medical equipment from other devices such as mobile phones, Radio Frequency Identification (RFID) systems and microwaves. EMI could also emanate from magnetic fields of other medical equipment such as dental equipment and magnetic resonance imaging (MRI). This research employs a wireless sensor network to monitor the electromagnetic compatibility of medical equipment by detecting the amount of radiation in its vicinity and predicting the possibility of electromagnetic interference using a fuzzy neural network approach. To detect the electromagnetic signals, a Radio Frequency meter equipped with low-frequency and high frequency probe, connected to an ESP32 microcontroller is used. The collected data is sent by Wi-Fi to the cloud for storage and analysis. A fuzzy neural network is introduced to the EMC/EMI monitoring to determine when electromagnetic signals are likely to be at their peak values with a chance to interfere with the normal operation of other equipment. The findings have shown that EMI prediction using ANFIS can be performed with high accuracy. The trained model demonstrated an excellent prediction capability when evaluated using test data, with RMSE of 0.334172, R2 of 0.9636 and MSE of 0.180645. The study aids biomedical technicians and engineers to manage the electromagnetic compatibility of medical devices and improve overall hospital safety to both patients and staff.
Keywords
Performance evaluation, Wireless sensor networks, Medical devices, Electromagnetic interference, Predictive models, Fuzzy neural networks, Electromagnetic compatibility
Rsif Scholar Name
Chiedza Hwata
Thematic Area
ICTs Including Big Data and Artificial Intelligence
Africa Host University (AHU)
University of Rwanda (UR), Rwanda
Recommended Citation
Hwata, C., Rushingabigwi, G., Gatera, O., Twizere, C., Mukanyiligira, D., & Thomas, B. (2023). Predicting the Risk of Electromagnetic Interference in Hospital Equipment Using Fuzzy Neural Networks. 2023 Photonics and Electromagnetics Research Symposium, PIERS, 1389-1396. https://doi.org/10.1109/PIERS59004.2023.10221355