Andriod tablet EFIS

[oahupilot]

Finally got some parts in for tablet display, namely the IOIO board, which allows me to bring in inputs to the usb port of my tablet. The ioio board is really cool you can read in analog, send pwm, read in or send out digital. You can use the board for really a lot of stuff.
here is vid of an rc car moving around only guided by the andriod phone alone linked to the ioio.
youtu.be/-H6MtevUjBg
The next bit of chip I am looking at comes from www.vectornav.com, Its better than the stock mems meters in consumer electronics because of its filtering. Its pricey so I might have to think about it for a bit.
youtu.be/amoFIo4MgT8
If any of the members on the list are EE majors or know how to code and design electrical stuff feel free to join in on the experiment. Sorry this won't be programmed on punch cards for you older folks.

[hans]

Oahupilot,

thanks a lot for these links. The second one jumpstarted a few ideas with which I will likely solve a programming/math issue I have been battling with...

Thanks again
Hans

[oahupilot]

Jan 25, 2012 13:29:50 GMT -5 hans said:

Oahupilot,

thanks a lot for these links. The second one jumpstarted a few ideas with which I will likely solve a programming/math issue I have been battling with...

Thanks again
Hans

The kalman filter?
I think their are scripts for it in matlab

[hans]

there's more than just Kalmann filtering going on.

Note in that video how (with the ahrs stationary) magnetic interference is NOT showing up on the right screen.
Apparently they have built separate state models for both the accelleraton sensors and the magnetic sensors, each model validating the other.
Like in: if my accelleration sensor model doesn't show any rotation in any direction, I deem any magnetic rotation to be caused by magnetic disturbances.
Or like in: if my accelleration sensors show a jitter which is not confirmed by the magnetic sensors, smooth out the accelleration sensor with this data.
Deriving a semi-complete movement- and position-model from each of the sensor sets and using these to validate the other one is a really smart thing to do. One requirement to achieve this would be to use sensor sets that both have a sensitivity that allows it to be used as a filter for the other sensor set.
It adds to the Kalmann filtering that likely will be smoothing out the signals coming from both sensor sets already...

At least that is what I got from the video...

Hans

[oahupilot]

I think you might be confused on what a kalman filter is.

Kalmann filtering is the statistical approach to determining the future position or state of something based on prior history and current information. Its not a hardware filter its a software filter.

I think this video helps explain a bit better what kalman filter is really.
youtu.be/86UeUvI7pLQ

[hans]

I know what a Kalmann filter is, had to program one myself a couple of years ago ;D


cheers
Hans