Analyzing Multimodal Sensory Signals Using Unsupervised Machine Learning

The Internet of Things (IoT) relies on sensor technology to enable seamless communication and data gathering across diverse devices and systems. This study explores the functionality of flexible and multimodal sensors based on GNP/HEC nanocomposites. These sensors hold promise for a wide range of applications spanning from medical contexts to industrial settings. A central aspect of comprehending IoT data organization is clustering analysis. In our study, we contribute by integrating GNP/HEC sensors into an Arduino system and applying the k-means clustering algorithm to identify patterns and anomalies in environmental monitoring. Our evaluation demonstrates that optimized clustering with k = 6, resulting in a Silhouette Score of 0.6557769, has the potential to significantly enhance industrial safety applications, particularly in the monitoring of temperature and ammonia levels.