Follow

# How do I calculate position and/or velocity from acceleration and how about integration drift?

The MTi’s (MTi-1, MTi-10 and MTi-100 series) and MTw have been designed as orientation sensors. The MTi-7 and MTi-G-710, although also primarily orientation sensors, offer velocity and position output and can be used in outdoor applications.

The inertial data outputs of the MTi or MTw can also be used to compute Position or Velocity data. Examples include estimating ship motion in the maritime industry

If you intend to calculate velocity and/or position, from acceleration and orientation data, using the MTi or MTw, you will need to subtract the gravity component from the acceleration data and provide an initial velocity. Double integration of acceleration data to estimate position is very inaccurate due to integration drift inherent with sensor noise and bias. When integrating sensor data, you will implicitly have to accept integration drift. Expect the position estimate to be acceptable for a short period of time only, in the order of seconds.

In order to get the free acceleration the gravity needs to be subtracted in the global frame. The calibrated data from the MTi /MTw however is expressed in the object frame. The object is the Motion Tracker (MT). To convert the measurements to the global frame, the orientation output of the MT can be used.

Please note that you need to use either the quaternion or matrix representation of the orientation for that purpose. Euler angles are not suitable for this kind of calculation. By using the Free Acceleration provided by the MTi-2, MTi-3, MTi-20, MTi-30, MTi-200 or MTi-300 you can skip the gravity subtraction. This is already done by the filter.

So called "zero velocity update" techniques allow to minimize the drift in situations where people are walking. Refer to this article for an example. For vehicles, the combination of MT data with other aiding sensors, such as odometers, flow sensors or cameras, is often used for simultaneous localization and mapping (SLAM) algorithms.

==

Disclaimer: be reminded that the MTi and MTw are designed as orientation sensors. Xsens supplies the Free Acceleration output and acceleration output as is. This article is published for courtesy reasons. As we are bounded by the physical limitations of the MTi and MTw, we cannot support you in calculating position, nor can we provide example code for this purpose.