The calibration required for the grab exercises required calibrating the G3. Doing this in coordination with the application was complicated by the fact that the communication between the G3 and the app was one way. This means the app has no way of communicating it's state to the glove. Thus, it was decided that they would operate off synchronized timing. Upon entering calibration, the patient was prompted to hold their hand open. 10 seconds later, they were prompted to curl their fingers into a fist. 10 seconds later, they had to pinch the thumb sensors. 10 seconds after this they had to pinch the palm sensor and calibration would end after 10 seconds of this. At the end of each 10 second window the glove would take a sample form the sensors to calibrate the min/max flex, and the min/max force the patient could generate.
Most of the grab based exercises had lots of overlap with the table top exercises. The exception to this was exercise G, H, and I. Exercises H was simple as it only required the user to squeeze the force sensor in their palm and hold it at max force for 5 seconds. This approximated the push-button light test in a direct way. Exercise I was less exact as it required the patient to use both hands to remove a lid from a virtual box. The box was anchored to their left hand and they would grab with their right. This served to approximate the Tupper wear test as it required both hands to work on the same task and for the patient to maintain a grab while moving their arm. Exercise G was the hardest to approximate as it required tearing paper. We determined this could be approximate by testing the pinch strength and seeing if they could maintain it while moving their arm in a downward motion. The final result was the patient being prompt to grab and apple from a tree and maintain a pinching force while pulling it down to their lap.
Exercise I
Exercise G