Acceleration on a Rollercoaster (And Integration to Velocity)

[DLDude]

Hey everyone. I'm working on a project that will incorporate a wireless accelerometer on a model rollercoaster. My issue is that we wish to take the accelerometer data and integrate it to give us our velocity through any part of the ride. This in turn will help us determine losses, etc. My issue is that the rollercoaster has loops and hills. It would be easy to integrate the 'X' accelerometer data on a flat plane, however since the coaster will be angled up or down depending on the part of the ride, you'd have to consider the affect of gravity in the 'x' direction when the car is tilted. This can be done easily with vectors by hand, but we need to be able to integrate in real time using pure accelerometer data (which is in normal values such as 'x=1.00 y=.40'.

Attached is a photo explaining my axises.




I know the Wiimote does something like this. My guess is we'd need to determine the angle of the bank and then take out the gravity aspect. That could become hard because through a loop, the Y axis data will be very large (along the lines of 5-6G's), so you cannot just compare the X and Y data and do some math there.

Any ideas?

 

[rcgldr]

The issue is that you need some known initial conditions so you know which direction gravity is relative to the accelerometer. Then you'd hope that your integration (perhaps numerical) is accurate enough so that the calculated results very close match the actual results.

 

[DLDude]

Initial height could be pretty standard, plus it will begin horizontally. I have a hunch you mean there would have to be a pre-set track configuration to match the graphed results to.

 

[rcgldr]

Initial height could be pretty standard, plus it will begin horizontally. I have a hunch you mean there would have to be a pre-set track configuration to match the graphed results to.

The configuration information isn't needed the accelerometer and integration techinque are accurate enough, but it would have to be very accurate.

Note that internal inertia based guideance systems were used on early long range missles, so high accuracy is possible, but I don't know how accurate your accelerometer is or the effect of accuracy versus distance or time varies versus accuracy.