Data Collection and Processing
Today Mira and I worked to create a more accurate, concrete algorithm to determine the difference between trembling and other movements that our patient does with her hands throughout the day. Our original program was based on numbers that we estimated to be decent cutoff frequencies to detect the kind of movement we were looking for. We realized after collecting data from the patient’s movements yesterday that there were certain movements, such as excitedly using hands while engaging in conversation, that fell within our criteria for trembling with our original code. We decided to look at the range in amplitude of the motion recorded from the accelerometer. After collecting this data, we were able to come up with additional constraints to make a more accurate algorithm. By increasing the amount of data we collected, we were able to more clearly detect the type of trembling motion we were looking for and decrease the chance for false positives.
-Kate
BME Design 