The key must be in having the greatest corrispondence of sensor values per the templete and signature manipulation, and/or the classification errors as to weather to have to creatfe a new template; ie ''While not required in all aspects, the training and updating in operations 330, 335 could be performed using Fuzzy Adaptive Resonance, Learning Vector Quantization, or other techniques whereby a computational device learns from
detected errors in order to improve future reactions to like sensor inputs.'' lol This stuff is way beyond me and anyone of ordinary skill IMO.

I can conceive of many ways the everday unique signitures could be utilized to verify identity, therfore allowing higher security, to allow a secure transaction to occure, take the UWB ablity of the 522 patent; it could simply take the configuratiion of the inside of your bedroom you woke up in and a few other inique sensor values that are only particular to you; such as something as simple as your coffe pot settings, and your wifes at home movement values and combination of you/her's breathing signatures when in bed, lol, even to your median driving route from house to work, or even look at you addrss book in your phone; grab a few voice values n match them to ph numbers...
The possibiliries are endless, and exactly what Intel was invisioning several years ago; I put up a great read on it per securityin the work place usinkg daily social sensor data gathered from multiple sources to identify you.

Ergo., ''In one embodiment, when enough events indicative of a particular user social activity are detected,
the calculating logic 150 identifies the activity as being performed by the user.''

emit...