Skip to main content
RACELOGIC Support Centre

VB3i IMU Kalman Filter



VBOX 3i has the capability of integrating the GPS data with inertial data from the IMU03 or IMU04 Inertial Measurement Unit in Real-Time, allowing accurate filtering and smoothing of the following parameters:

  • Latitude
  • Longitude
  • Velocity
  • Heading
  • Height
  • Vertical Velocity

The advantage of VBOX 3i - IMU integration over non-IMU Kalman filtering is that the Kalman filter is using physical inertial measurements from the IMU and GPS engine together. This allows it to rely on IMU data when GPS signal quality is bad or absent (for a short duration).

Noisy and intermittent GPS Signal

The Racelogic Kalman filter is capable of using GPS measurement data and IMU inertial measurement data together to compensate for GPS dropouts or noise. Below is a plot of the number of satellites (blue) and GPS velocity (red) which is experiencing noise then dropouts due to the vehicles environment.

1a (1).png

The plot below illustrates the difference in quality between the Kalman filter derived velocity (black) compared to the velocity derived from GPS alone (red).

The Kalman derived velocity can be seen to be a very accurate estimation of the velocity of the vehicle despite the adverse conditions.

2 (1).png

To illustrate the accuracy of the Kalman derived velocity, below is the the Kalman derived velocity in black and the wheel speed derived velocity in blue.

3 (1).png

Long duration dropouts

The plot below shows a complete GPS dropout of approximately 6.5 seconds. 


4 (1).png

Comparing the Kalman derived velocity (black) to the Average Wheel speed (blue) you can see that even with no GPS input the Kalman estimate is very good when relying solely on IMU data.

5 (1).png

Hardware configuration

Please see the VBOX 3i manual 'IMU Integration' page for details of how to configure hardware.

Firstly, VBOX 3i needs to be configured to use the IMU data it is connected to.


The VBOX Kalman filter module is designed to take a standard .vbo file that has been logged under poor conditions (for example in close proximity to trees and buildings) and from this generate another .vbo file with more accurate speed and position data.

The reason the Kalman filter can improve raw data is because the position and speed are measured using two different methods in the original .vbo file; speed is measured using Doppler shift, and position is measured using normal GPS trilateration. Speed and position are related very closely, which means you can use speed to calculate position, and vice-versa.

As an example of the position smoothing, the chart below is a circuit outline taken at the Nurburgring. In this file, trees close to the edges of the track have caused poor satellite reception resulting in noise and steps in the data.

The original file is indicated by the red trace and the Kalman filtered version by the blue trace. As can be seen in the blue trace, the Kalman filter has correlated positional and speed data to remove the steps in position without the loss of data that would occur with other filtering methods.

10 (1).png 11-2.png

The speed data shown on the right shows the noisy original in red and the Kalman filtered in blue.

Note: If you are measuring braking distance, the Kalman filter must not be used.




These are Kalman filter diagnostic channels used to help Racelogic troubleshoot and improve the Kalman filter in different environments/ conditions.








Channel specifically used to track the synchronisation between the IMU and VBOX.

Note: Covariance and T1 channels are calculated in the .vbo file by default when IMU kalman filter integration is enabled.

  • Was this article helpful?