​What is soft tissue artifact
Although the name may seem a little scary to novices, the soft tissue artifact is the biomechanical engineer's perennial nightmare. What is it about?
Briefly, it is a fictitious motion effect that a sensor can experience when placed on the body near contracting muscle parts or moving skin.
​
In the previous sections we had seen that the information provided by the IMU is related to the movements perceived by it. Therefore, the accuracy of this perception depends on how firmly the unit remains on the body part it is measuring.
​
​
Example
To give a practical example, our arm is surrounded by very different muscles that change their shape based on the movement performed (after all, it is an integral part of their work!). By rotating the forearm upwards, we use our biceps.
Therefore, during a movement like this, if the IMU measuring the distal segment were positioned above the biceps and the IMU measuring the proximal segment were positioned on the forearm, the movement of the elbow would still be calculable, but this calculation would lead to a result that changes during the rotation of the elbow itself, as the IMU on the biceps would be subject to fictitious rotations caused not by the movement of the underlying arm, but only by the contraction of the muscle.
The forearm IMU could at this point also measure with 100% accuracy, but since the calculation involves measuring how much the forearm IMU rotates relative to the arm IMU, this relationship will be subject to errors and will not lead to a truthful representation of elbow movement.
​
​ concept is not something that can be overlooked during the development of a movement analysis system or protocol, as the error generated by the artifact can reach very high percentages of the actual measurement. In fact, try to imagine the case of a very muscular person: the IMU on the biceps could even bend laterally as if the arm were rotating 90 degrees...
How to deal with the problem?
From a mathematical point of view there are attempts to remove this error by filtering “in the frequency domain”. It's a shame that this error, being linked to the muscles that generate movement, is linked as dynamics to the movement itself. Therefore filtering this error is equivalent to removing information on the movement itself.
The solutions used over the years are varied and, although they can be classified based on good performance, the soft tissue artifact is not something that can be completely eliminated or compensated for.
If on the one hand particular bandages or other more mathematical stratagems can be used, the technique that Turingsense has used over the years concerns the use of particular areas of the body segments for the good seal of the IMU on the body and the maintenance of a very low level of artifact. Naturally this method has involved numerous studies and research.
​
For more information on our research, visit our bibliographic section.
