Ruprecht-Karls-Universität Heidelberg
Optimization in Robotics and Biomechanics

Oberseminar 01.12.2014, 11h c.t. (H2.22)


Practical Human Movement Prediction: The Missing Pieces


Dr. Matthew Millard, Universität Duisburg-Essen


Computational biomechanics is changing the way basic research is undertaken, how products are designed, and how surgeries are planned. Despite all of the progress, predictive biomechanical simulation methods can be improved. The simulation methods used to estimate the muscle forces and tendon strains of a subject during a recorded task are time-consuming and error prone. The methods used to find the optimal movement of a musculoskeletal model (with elastic tendons) are also very resource intensive - requiring 100's to 1000's of CPU-hours to converge. Fortunately there is an excellent method from the field of dynamic bipedal robots that can be used to bring practical prediction to the scientists, engineers, and surgeons of tomorrow: hybrid zero dynamics. All that is required to apply this method to a musculoskeletal model is a holonomic muscle model, and a holonomic foot model. In this talk I will introduce the problems of muscle force prediction and movement prediction, and a potential solution: a hybrid zero dynamic musculoskeletal model. 

K. Mombaur,
Last Update: 26.11.2014 - 13:26

The photographs in the header of this webpage have been taken at the Musee de l'Automate in Souillac, France