Hmm... so should I put you guys down for zero sensors, or infinite "probes" placed all throughout the nervous system, recording the neuron pulses?
I confess, using some hyperbole in the thread subject, as my intentions were not so ambitious as to capture a feeling, whatever that may mean. Initial thoughts would be (assuming a data capture implementation):
Remembering static postures: how to remember that one body re-position from the visiting instructor that made all the difference but is foreign and awkward (like all new movements) and tough to remember.
Next being able to dynamically set ranges with some audible or vibration feedback whether thresholds are exceeded... moving through a waza without the shoulders raising... whatever the bad habit being targeted at the moment.
Basic stuff. Just finding or making the data visualization tools will be the biggest challenge. The data capture seems straightforward enough with off-the-shelf components streaming to a laptop or phone.
Prototyping a solution once the parts arrive:
should be enough to get the ball rolling, so to speak.