Modeling human behavior from wearable sensor data to improve health and healthy aging.
Meet the team See our publicationsWe are a team of engineers and researchers working to have a positive impact on people's health through pervasive sensing and machine learning that lasts a lifetime.
Engineering research
Machine learning methods for activity recognition and physiological sensor data processing, connecting behaviors and health.
A focus on people
We study how researchers and end users interact with our methods, and how they shape real health decisions.
Great tools
Wearables, physiological sensors, and data pipelines for collecting, storing, and processing large-scale sensor data.
Interdisciplinary research
We collaborate across engineering, psychophysiology, and clinical fields on problems related to health and aging.
Our current research program (2026–2031) pursues two aims: making models robust to data heterogeneity across devices and sampling rates, and building interpretable models of long-term behavioral dynamics.
A reusable, service-oriented pipeline for ingesting, harmonizing, and processing heterogeneous wearable sensor data, built to make precision-healthcare research reproducible across datasets, devices, and sampling rates.
Physics-based simulation and data-augmentation methods (WIMUSim) that make IMU-based human activity recognition models robust to sensor placement, orientation, and measurement variability.
A Bi-Mamba backbone with a Triadic Cross-Attention module for predicting sleep stages from wearable PPG signals, extending the lab's earlier SleepWatchNet, evaluated against public PPG and accelerometer sleep datasets.
The lab was awarded a MITACS Accelerate grant (Sep 2026 – Dec 2027) for “Autonomous Multimodal Knowledge Graph Reasoning (QSIG) for AVACS™-Driven Human Digital Twin and Edge Clinical Autonomy.”
“Physically Plausible Data Augmentations for Wearable IMU-Based Human Activity Recognition Using Physics Simulation” was published in IEEE Sensors Journal. Read more →
Dr. Lago was awarded an NSERC Discovery Grant (2026–2031) for “Robust and Interpretable AI for Modeling Human Behavior in Real World Contexts,” funding the lab's two main research threads: robustness to data heterogeneity, and interpretable models of long-term behavioral dynamics.