ID and Joint Reaction inconsistencies with the applied External Loads

[Emanuele Sessa]

Hallo,
Im testing a low-back exoskeleton similar to CrayX (https://www.germanbionic.com/en/crayx/), with the difference that CrayX has a lower module connected to the thighs while my model is attached only to the pelvis and the shoulders, with a motor between the pelvis and shoulder modules.

My interest is in the effect of the exoskeleton on a stoop lifting routine, so I first ran the analysis on a full-body model without the exoskeleton in order to obtain the base values of forces and torques in the pelvis-low back area, and then added the exoskeleton to the model and ran the analysis again.

In order to account for the action of the exoskeleton, I imposed a certain torque pattern for the motor and computed in MATLAB the corresponding time profile of the force acting on the shoulder module (which pulls the shoulder during the lifting); then, I added such force in the External Loads file, expressing both the force and its point of application in the reference frame of the torso.

It must be noted that the torque I imposed to the motor is active also during the phase of trunk flexion, so that in every time instant of the routine the torso of the user experiences a torque that tries to pull it back to the upright position (just like a rotational spring). This results in a force at the shoulder module that is always pointing backwards with respect to the frontal plane.

Moreover, I added in the External Loads file also a torque acting on the pelvis, opposite to the one imposed to the motor in order to ensure local equilibrium of the actuator.

I used such External Loads file to run an ID, a SO and a Joint Reaction analyses, and compared the results to the base values previously obtained.
The ID comparison results, shown in the picture attached, explain my doubt: the moment around the L5-S1 joint is always smaller for the routine with the exoskeleton, but I would expect it to be higher during the trunk flexion phase (15s-18s and 22s-24s) since then the action of the exoskeleton is contrasting the motion, and smaller only in the lifting phase (18s-20s and 24s-26s), as then the exoskeleton is supporting the motion.
I checked the same quantities in the results of the Joint Reaction Analysis, and again found the same relationship between the profiles.

So my question is: why isn't the relationship between the profiles of the moment changing with the phase of the routine? What exaclty comes out from the ID, if not the generalized torque that the muscles have to overcome in order to induce a motion?


I will be extremely thankful to anybody who could give me a hint. I hope this can happen in the shortest time possible, as unfortunately I have a very tight deadline for this project.

Thank you and best regards

Emanuele