The problem
Blood pressure is one of the most clinically useful signals in medicine and one of the most poorly sampled. A cuff gives you a handful of readings a day, taken in the least representative environment imaginable. Continuous, noninvasive estimation has been a standing challenge precisely because the signal is not directly observable — it has to be inferred from a set of correlated biosignals, each with its own noise, drift, and failure modes.
What we built
A novel multisensor wearable system capable of noninvasively collecting a suite of biosignals and biomarkers, paired with AI models that estimate blood pressure continuously. The system is distributed and networked rather than a single sensor-on-a-wrist, which is what makes continuous estimation tractable.
The engineering problem sits across every discipline RNS practices at once:
- Sensing — capturing multiple biosignals cleanly enough to be useful
- Embedded systems — doing it at a power budget compatible with continuous wear
- Signal processing and AI — turning correlated, noisy inputs into an estimate worth acting on
- Wearable design — packaging it so a person will actually keep it on
Outcome
The work produced a pending patent and multiple publications, and is detailed extensively in the thesis Continuous Health Monitoring with Distributed Networked Wearables.
Why it matters here
This is the archetype of the problem we take on: a device that only works if the mechanical, electrical, firmware, and model decisions are made together. Split that work across four vendors and the integration is where your runway goes.