I am trying to figure out how the contact conditions in the seated human model actually work. From my understanding of the description, contact nodes are defined, so that the contact between the body and the environment segments (back rest, seat, etc.) are only active when the distance between the body contact nodes and the environment segments is less than 5 cm. Is that correct?
In this case, how does the contact work when it is active? If the distance between body node and support is less than 5 cm, is a force generated so that any movement towards the support surface is not allowed? Or is the contact force a function of the distance between the contact nodes and the support surfaces?
For example, for a subject lying horizontally (e.g. bed rest), if the spine contact nodes are close enough to the back support, how does the contact work? I see some spine flattening, so I guess the vertebrae move closer to the support even if the contact is active, is that true?
Sorry, I got confused, the vertebrae do not seem to get closer to the back rest. However, is it true that when the contact is active there are reaction forces that prevent any movement towards the backrest? And that only sliding (governed by friction) is allowed?
Also, since the backrest has a thickness (and thus a upper and lower surface, when the human is lying horizontally), what is, exactly, the surface of the backrest used for the contact?
i will try to explain a bit more about the contact elements.
The contact conditions in the seated human is actually an array of artificial muscle that allows forces to be transmitted when the points are close enough. It also linke normal force to frictional forces, so it does not allow a larger frictional force than the normal force can hold, if can be less of course.
thank you for your answer. I still don’t really understand: how are the contact forces calculated? I though the base object was something like a ground which cannot be penetrated and that contact forces were basically reaction forces that prevented penetration between base and target.
Is that correct? If so, is it true that there is no relative movement in the normal direction between base and target (except if their distance is increasing as I read that contact forces only push, don’t pull)?
Also, you mentioned that the contact is between points, not surfaces. In the code of the file BackrestSupport.Any I see that the Base Object for every spine node support (T9Support, T10Support, etc.) is Seg.Backrest, which is not a point, but a segment, correct?
In this case, I can understand what are the target objects, as they are Support nodes, but what is exactly the base object?
The contact forces element do not prevent any motion there are no motion drivers associated with these elements, so this will not stop motion, but they can generate forces.
The base object needs to be a refframe so this can be either a AnyRefNode or simply as you have seen an AnySeg, both are a refframe.
The word baseobject is just a name it could have been anything it was named like this because it is the objects which “supplies” the force.
The contact force are calculated as part of the muscle recruitment, since these elements are actually made using artificial muscles. So what happens is kinematics is resolved first, then the muscle recruitment try to see how the boundary conditions provided (contact elements) can be used to create balance in the model. So the muscle recruitment activate the contact elements to create balance and to unload the real muscles in the model. The contact elements has been constructed to incorporate friction, the coef is a constraint that ensures a balance between normal force and frictional forces, without this constraint the model may choose to have frictional forces and no normal forces.
If you need a setup where there is connection between force and motion then force dependent kinematics is needed. Using force depending kinematics motion can be “stopped” by contact forces arising from elastic stiffness defined as surface surface contact or simple springs.
thank you, it is clearer now. Just to confirm, then, if I simulate a subject lying horizontally or seating on an inclined backrest (using the Seated Human model), and I do not define any relative motion (using joints and drivers) between the human and the backrest, I will not see any relative movement between base and target, correct?
And if I get it right, the backrest segment that I see in the model is only for visualization purposes, but the contact is actually defined between the contact nodes and the backrest reference frame, is it? Then the contact volume is defined by UserDefinedLimitHigh, UserDefinedLimitLow, and UserDefinedRadiusLimit? So basically it will be a cylider, instead of a rectangular-section shape like the one that visually represents the backrest?
If there is no motion defined between spine and backrest it will not move. In the seated-human model there is a driver with constant distance between backrest and thorax.
Concerning the contact elements you are also correct on that this is cylinders not boxes, the box is just for visual purpose.
just another very quick thing: is there a published paper reporting the validation of the seated human model, by any chance? I found a webcast, but I couldn’t find a published paper…
Please have a look at this paper, it is a somewhat different setup than the seated human, but the principles are the same.
[ol]
[li]Olesen CG, de Zee M, Rasmussen J (2014), “Comparison between a Computational Seated Human Model and Experimental Verification Data”, Appl. Bionics Biomech., vol. 11, pp. 175-183. [DOI, WWW]
[/li]
[li]Olesen CG (2012), “The Influence of Sitting Conditions on Soft Tissue Loads”, Thesis, Aalborg University, Denmark. [PDF]
[/li][/ol]
