I have always used Kalman Filter to smooth a signal comes from one sensor only.

I would like to know if Kalman Filter could be used to fuse data coming from two different sensors that provide the same type of reading. An example is fusing the position data that comes from the GPS with the position data that is calculated from the IMU. In the light of this example, I wish to have the following:

  • State vector: [x;y]
  • measurements vector: [x_imu;y_imu;x_gps;y_gps]

Let me know if you need more information/clarification.


1 Answer 1


The answer to the first question is yes, it is possible to fuse data coming from the same type of sensor. For example, the fusion of two different accelerometers, each with a measurement quality, gives a weighted average where the weights are the inverse of individual variances (see Kays's[1] book for details).

The answer for the second is yes as well, it is possible to fuse IMU with GPS measurements. But note the one returns a position itself, while the other measures the second derivative of the position. The fusion demands well elaborated models. There are plenty of papers and books with models and its theory. I would recommend, for example, Farrel's[2] and Titterton's[3] books.You can easily find implementations of the fusion filters in internet. Open source projects of drones, for example, have it.

[1] KAY, Steven; Fundamentals of Statistical Signal Processing, Estimation Theory.

[2] FARREL, Jay A.; The Global Positioning System and Inertial Navigation.

[3] TITTERTON, D. H.; Strapdown Inertial Navigation Technology.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.