the problem
Falls are among the most damaging and most predictable events in patient care. The predictors are in how a person walks and in physiological signals that are recorded routinely — but rarely analysed together, and almost never in advance.
what we built
Predictive algorithms that combine gait analysis with physiological signal processing to estimate fall risk. The work covers the full chain: extracting gait features from sensor data, processing the physiological channels, and fitting models that produce a risk estimate a care team can act on.
what it changed
Fall risk becomes something that can be scored ahead of the event, from data already being collected, instead of reconstructed after it.