Xsens Releases Algorithm for Stable 3D Tracking in Magnetically Distorted Environments
July 7, 2010

Xsens Releases Algorithm for Stable 3D Tracking in Magnetically Distorted Environments

Enschede, Netherlands - Xsens Technologies B.V. has announced the general availability of its KinematicCoupling (KiC) algorithm in its newly released software suite, MVN Studio 3.0 for professional grade human motion capture. The KinematicCoupling (KiC) algorithm is the latest addition to the advanced signal processing suite developed by Xsens for processing inertial sensor data (accelerometers, gyroscopes, magnetometers) to accurately capture motion in 3D.
Until now, adoption of inertial sensor technology for human motion capture was hindered due to the fundamental dependence on the Earth’s magnetic field to provide a sense of horizontal direction (heading) for long term measurements, much like how a compass works.

The patent-pending KiC algorithm -- the culmination of years of research at Xsens -- solves this problem by taking a new approach, with the result of total immunity to magnetic distortions when estimating full 3D joint orientations. “Using KiC, it is now possible to accurately track movement in environments with significant amounts of iron, like large steel beams, ferroconcrete floors and radiators,” says Per Slycke, CTO of Xsens.

The advent of the KiC algorithm is a corner stone to build on to provide robust, accurate 3D tracking of human motion under all circumstances, fully establishing inertial sensor technology as a disruptive technology to be reckoned with in the 3D tracking field. The Xsens MVN full-body tracking system, which is already in use among top developers at Electronic Arts, Industrial Light and Magic, Sony Pictures Imageworks, Daimler, Össur, INAIL and many other companies and Universities, benefits immediately with the 3.0 release of MVN Studio.

KiC inventor and research manager at Xsens, Dr. Henk Luinge (PhD), explains: “A system with (bio)mechanical links, such as a human being, contains a lot of correlation between the kinematic data measured at each point on the body. Basically, by comparing the acceleration, measured by each sensor, in each joint, we can directly determine the full 3D orientation of that joint without the need for magnetometers.” An additional important advantage, according to Dr. Luinge is: “In the first release of MVN Studio 3.0 the position of the inertial sensor with respect to the joints must be input to the algorithm. But, we already know that it is possible to estimate this position accurately based on the data gathered during normal use of the product. Essentially, this will enable us to always get the best calibration possible, for each individual user, automatically.” These latest developments are currently being researched.