University of Konstanz
Graduiertenkolleg / PhD Program
Computer and Information Science

Guest Talks


Applicability of functional joint axis calculation: implementations accounting for the kinematics of the intact human knee


Dipl. Ing. Dr. techn Irene Reichl, University of Vienna, Austria

date & place

Tuesday, 17.04.2012, 15:15 h
Room Z 613


The analysis of human movement relies on a reliable determination of joint rotation axes and centres. Especially high accuracy is required in the field of high performance sports as misaligned axes modify the kinematics and kinetics calculation. Furthermore, in clinical applications a reliable determination of joint rotation axes and joint displacements is crucial, since the kinematical analysis should assist surgical interventions or prosthesis design. These axes and centres may be estimated from predictive formulas based on anatomical landmarks. However, for some joints more than one definition is in use, and the landmark palpation is observer dependent. Alternatively, the relevant reference axes are functionally obtained from kinematical data of a well-defined type of movement. Here, a mathematical optimization process based on an appropriate kinematical model of the joint in question yields the requested joint axes or centres. The objective of this study is the investigation of mathematical procedures with special emphasis on the human knee joint. However, measurement errors and noise render the calculation quite difficult. Those are expected to be compensated for by kinematical constraints reducing the number of degrees of freedom (DOFs), thus, yielding an increased reproducibility of the joint displacements. The simplest model of the knee would be a hinge, i.e., a joint with one rotational DOF. As the immediate more complex model, the finite helical axis permits translation and rotation (2 DOF) about the same axis. However, an approach that accounts for the specificity of the healthy human knee, approximates the joint as a compound hinge joint (CHJ), a model which has been already successfully applied by several groups. This approach assumes that the knee can be approximated by two rotational DOF, only, namely flexion and tibial rotation. The other DOF are minimized. Within the present study, common implementations of the CHJ model are examined and modified by changing the number of model parameters. On one hand, the geometric constraint between flexion axis and tibial rotation axis is released, on the other hand, the algorithm identifying the flexion axis is modified. The conventional and modified implementations of the CHJ model are applied on fictive kinematical data in order to compare their effect on the obtained displacement. Since increasing noise distorts the shape of the objective function, the determination of its minimum becomes more difficult. Thus, it can be concluded that the decision for a specific model should be based on the data at hand. Kinematical data from medical imaging or bone pins in cadaver data permits a description involving a larger number of free parameters and a refined model. In contrast to that skin-mounted markers involving soft-tissue artefacts demand a more constrained model.