You are here: Welcome » m3_seminar » M^3 Seminar 2019 » projects_ustutt » Equilibrium-Point Control

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision		Previous revision
m3_seminar:m3_seminar_2019:projects_ustutt:ep-control [2019/07/16 10:49] – created emjaym3_seminar:m3_seminar_2019:projects_ustutt:ep-control [2019/07/30 22:55] (current) – [Equilibrium-Point Control] stud_ustutt
Line 1: Line 1:
-EP-Control+====== Equilibrium-Point Control ====== 
 + 
 +^ Autors | Pohl Steffen & Manglkammer Philipp | 
 +^ Betreuer | Marc Jacob | 
 +^ Bearbeitungsdauer | ca. 25 Stunden | 
 +^ Präsentationstermin  | dd.mm.yyyy | 
 + 
 +<note tip> 
 +You can find more information following the next links: 
 +  * Wiki page ** formatting **: [[https://www.dokuwiki.org/wiki:syntax#formatting_syntax|HowTo - Wiki-Format]] 
 +  * Creation of a **video-podcast ** [[http://wiki.ifs-tud.de/howto|HowTo - Podcast]]  
 +</note> 
 + 
 +===== Introduction/Motivation ===== 
 +Basic EP-concept (EP = Equilibrium point) 
 + 
 +- The EP-Control System states, that a set of two actuators can act like a human muscles Agonist/Antagonist principle. 
 + 
 +- The applied force depends on muscle length and stimulation: The longer the artificial muscle, the more effective the stimulation in translated into force 
 + 
 +- The system is heading towards a certain joint angle depending on the realation of force and length of both actuators 
 + 
 +- Once the angle is reached, the joint stiffness can be modified by equally increasing or decreasing the applied forces in the actuators.  
 + 
 +- In our experiment, the used software defines sitffness and angle of the models joints which are automatically are translated in poses. Each of these poses define an Equilibrium Point 
 + 
 +===== Podcast ===== 
 + 
 +Movement approaches via poses: 
 + 
 +- The first set of poses was designed to be as detailed as possible and utilized every possibly relevant angle 
 + 
 +- The second set only used simple movement patterns. The more complex movements then resulted from the passive mechanics of the muscle system 
 + 
 +Resulting movement: 
 + 
 +- the first set of poses result in failure due to loss of balance 
 + 
 +- the second set shows a good result even without specific modelling 
 + 
 +{{ :m3_seminar:m3_seminar_2019:projects_ustutt:biorobotikpodcast.mp4 |}} 
 +===== Summary ===== 
 +At the end, three conclusions were reached: 
 + 
 +1. The system is able to head towards a target position without the need for an inverted model 
 + 
 +2. Using rough and simple movement patterns looks more natural and is more stable 
 + 
 +3. A comparison between both approaches rises the question whether biological movement is also just a product of simple, prototypical poses instead of detailed motions 
 + 
 +===== References ===== 
 +  - Bayer, A., Schmitt, S., Günther, M., Haeufle, D.F.B. (2017): The influence of biophysical muscle properties on simulating fast human arm movements. Computer Methods in Biomechanics and Biomedical Engineering 20(8), 11-24. https://dx.doi.org/10.1080/10255842.2017.1293663 
 +  - Christensen, K. B., Günther, M., Schmitt, S., Siebert, T. (2017): Strain in shock-loaded skeletal muscle and the time scale of muscular wobbling mass dynamics.  Scientific Reports 7(1), 13266. https://dx.doi.org/10.1038/s41598-017-13630-7