Dead Reckoning precision requirements

[øysteinkristian]

Hi

We have a GPS product with NXP sensor MMA8652FC (acceleromoeter), FXAS21002C (gyroscope) and MAG3110 (Magnometer)

We want to use these sensors to perform dead reckoning if GPS coverage is lost. My guess is that the software algorithm for dead reckoning with these 3 sensors is some months development time and before we start this process I would like to do a preliminary calculation or educated guess on what level of dead reckoning precision we can from the last known GPS position if our GPS speed and GPS heading direction data was 100%correct at last know GPS position.

Can anyone say something roughly about the estimated preicison we might get here with these sensors?

Maybe 5 or 10% from distance traveled?

BR

Øystein

[michaelestanley]

I touched on this some time back in the posting at Position Reckoning .

That posting looked only at double integration errors associated with accelerometer offset.  There's also the issue or triple integration errors associated with gyro outputs.  Having said all this, you can probably improve performance above these numbers by running a higher level fusion function which incorporates your GPS data.