Dead-reckoning in the firmware?

I have been reading this code, as well as Bill's DCM paper, and I am wondering 
if there is an available implementation (of any kind) of dead-reckoning using 
the concepts from the DCM algorithm.

I noticed this file in what appear to be some of Bill's earlier implementations 
(from MatrixPilot): 
http://api.ning.com/files/OOZjvN9I7PtfdKmftHVruToOhl7IQQ4cfv*Pw8HbllP922hp7HimOb
7yIe*0hV0uVlqgmHKUHsDmwihbYvDFFg__/deadReckoning.c

It seems to provide dead-reckoning, but only when hooked up to GPS data.  It 
seems to me that the DCM concepts could be applied with a magnetometer to get 
precise relative coordinates in indoor situations.  Dead-reckoning could also 
be used to report precise positioning information in situations where GPS 
looses it's lock (ie a car's moment-by-moment turns going through a tunnel).

Are there any known implementations of this?  In general, are there any good 
open-source dead-reckoning algorithms I am missing?  It seems like DCM is very 
close to nailing down the position of an item (relative or absolute) instead of 
just nailing down its orientation.

Any comments on the topic of dead-reckoning would be greatly appreciated.

Original issue reported on code.google.com by benstand...@gmail.com on 22 Jun 2011 at 1:25

tsbertalan / ardu-imu

Dead-reckoning in the firmware? #16