Inertial Tracking, any idea?

vvvv beta question
1

Inertial System (INS) and Inertial Measurement Units (IMU) are relatively new devices (from '50s) for getting physical information about acceleration and orientation of an inertial object.
This approach leads to dead-reckoning situation, where no external device will adjust or correct the measured data. Dynamic Bayesian network and other algorithm are used to fuse data from accelerometer and gyroscope to make a smoothing and prediction of the changes, that in this case, would lead to an extreme error after a certain amount of space/time. (recursive error)
Several models exist, ranging from huge nautical to tiny wearable devices, the size of a small mobile.

Vmsens (CH) and Xsens (NL) provide pocket size solutio along with a complete SDK. Price ranges from 1000$ to 5000$ and up

2

found this cheap approach
http://www.sparkfun.com/products/10010

3

i saw a 9 dof today as well
i think the last 3 are magnetosphere readings (used with gyro = absolute rotation)

to integrate the data properly i presume you’d need to perform this in hardware or at least in a seperate thread (to ensure readings quicker than 60fps)

4

we were playing with XSens some time ago. Wasn’t much fun, it was bulky and not precise at all.

for that price you can buy a whole bunch of arduinos with wifi- and gyro-shields.

5

Tracking position in 3d real world space with IMU’s is a really challenging task, requiring much control over algorythm and sensor calibration. There are lot of Ph.D thesis concerning this issue.

6

anyone in to SLAM or PTAMM?