2025-CME-218

Deep Learning Based Analysis of Footstep Induced Floor Vibrations for Occupant Health Monitoring

Joshua Bruce Mehlman, P.E., Tsering Yangchen Yonzon

School of Engineering

Faculty Supervisor: Zhuwei Qin

We will apply advanced deep learning techniques and data processing methods to analyze floor vibration data induced by footfall. Specifically, we leverage successful image processing approaches by transforming vibration signals using the Continuous Wavelet Transform. We will experiment with a variety of standard wavelets, as well as a custom-designed wavelet. Insights gained from these experiments will be applied to both global and personalized models to extract detailed information about gait, with the goal of enabling early detection of health conditions such as Alzheimer’s disease.