I'm currently working with a huge amount of GPS tracks from vehicles. Looking at the data I noticed something interesting : I have a GPS position + timestamp every 10 seconds (which is not a lot) but I found out that vehicle heading and longitudinal acceleration are available every second in the data.

In your opinion what would be the best way to use these detailed heading and acceleration measurements to reconstruct the missing position between the GPS positions?

Kalman filter pop up regularly when working with GPS tracks but I'm not clear if it is only useful to smooth/clean the GPS tracks or if it could be of help here.

  • 1
    Bit of a vague question really, "opinions" aren't encouraged here! You'd have to build a joint model for location based on GPS and sensor data, modelling the correlation and so on. I don't think I can add more without actually seeing some sample data and how they relate. – Spacedman Sep 26 '18 at 13:28
  • If you have an underlying road network, the keyword 'map matching' might also be interesting for your research. In the open country map matching is, of course, quite useless. – tallistroan Sep 26 '18 at 20:05

Browse other questions tagged or ask your own question.