Summary
Indoor fungal growth can compromise air quality, building integrity, and human health, yet traditional identification methods rely on slow culturing techniques or manual sampling. This project developed Scensory, a robotic olfactory system capable of real-time fungal identification and spatial localization using low-cost volatile organic compound (VOC) sensors and deep learning models. The system integrates sensor arrays with automated robotic sampling to collect high-throughput VOC data from diverse fungal species. Neural network architectures were trained to classify species and infer their spatial origin from short temporal sensor signals, enabling rapid and accurate environmental assessment. This work demonstrates a scalable framework for autonomous fungal surveillance in built environments. By combining robotics, chemical sensing, and machine learning, the project advances real-time environmental monitoring and provides a foundation for early detection of harmful microbial contamination.