SmartCPR

Under the direct supervision of the President of the Royal Society of Canada’s Academy of Science I designed and patented a low-cost and lightweight prototype CPR mannequin outfitted with real-time objective feedback.

As part of McMaster University’s ongoing CPR initiative, promotional content for this technology was be translated into a number of languages to make it available to developing countries that stand to benefit the most from accessible CPR training technology.

The main challenge associated with using low-cost sensors to provide feedback on the depth and rate of CPR compressions is the inherent drift associated with the signal from inertial sensors like accelerometers, which is only exacerbated by the double-integration required to extract displacements.

The spectral analysis technique used to reconstruct a drift-free displacement signal from an accelerometer is shown above using real compression data (implemented in MATLAB).

To circumvent this issue, I identified a novel digital signal processing technique in a 2016 PLOS One publication that takes advantage of the periodic nature of CPR compressions and the fast Fourier transform (fft) algorithm to extract fundamental info from finite segments of sensor data that can be used to reconstruct a drift free signal in near real-time.

My implementation of this algorithm for our prototype formed the basis of a companion training app that was developled after my departure and is currently under developlment.

Early trials using the protoype mannequin and software were promising enough to prompt the disclosure of our work to the university patent office. As part of this disclosure process I worked closely with a medical doctor to draft a detailed patent claim & conflict analysis (as well as tentative distribution plans).