next-generation PMU discussion / Memory 1.8V

Andy Green andy at
Mon Apr 21 15:49:29 CEST 2008

Hash: SHA1

Somebody in the thread at some point said:

|> It's possible we might be doing the motion tracking for extended periods
|> if the guy is on the move or even in a car, it will show as exceeding
|> the threshold stuff all the while.
| Hmm, what do we actually do in this case ? Try to filter any gestures
| out of the car motion noise ? That sounds painful.
| Common wisdom is that you can't use acceleration sensors to measure
| large-scale movements (e.g., to enhance GPS performance), because of
| their poor low-frequency accuracy. So continuous integration while
| driving wouldn't make too much sense.

Yes dead reckoning reliability will be a problem.  But sometimes you can
do something with even low quality information that you definitely
couldn't target with no information.  For example, there's enough
kicking about to say "I'm travelling in vehicle", from the
accelerations, braking and sustained velocity behaviours.

Also, if we manage to get something working with the GSM "set" of
transponders, maybe we only ask the motion sensors to "fill in" the
position with a low level of accuracy until the next GSM sniff.

- -Andy
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora -


More information about the openmoko-kernel mailing list