# Extended Kalman filter (EKF)

I am working on the localization problem of an underwater vehicle. However the problem is still in very simple and it does not matter that it is about underwater.

How can I use the EKF when I have multiple measurements but with different sampling? Not all the measurements are available at the same time to perform the update step.

Can I perform the update for the measurements that I have at each time step? For example, I have a pressure sensor and a speed sensor with frequency $1$ Hz. But I have also another sensor (USBL) system with $0.01$ Hz.

What if I perform update step for pressure and speed sensor at each time step adjusting the measurement model to that case (like I have only these sensors) and once I receive the USBL measurement I augment the measurement vector including the USBL measurement?

• Any help? As far as I concerned, what do need is called: Data fusion (or Multiple sensors Kalman filter data fusion). – Gluttton Feb 15 '16 at 20:48
• This answer to a similar question might have some insight. Let me know if it does. – Peter K. Feb 15 '16 at 22:05
• thanks for the replies but still do not know how to deal with asynchronous observation. Assuming, for example, one has different sensors which provide the observation vector. However, each of these sensors provides measurements in different time steps. Then, who can I perform the update step of the EKF? – user19571 Feb 16 '16 at 13:06