A computationally efficient optimisation-based method for parameter identification of kinematically determinate and over-determinate biomechanical systems

Abstract

This paper introduces a general optimisation-based method for identification of biomechanically relevant parameters in kinematically determinate and over-determinate systems from a given motion. The method is designed to find a set of parameters that is constant over the time frame of interest as well as the time-varying system coordinates, and it is particularly relevant for biomechanical motion analysis where the system parameters can be difficult to accurately determine by direct measurements. Although the parameter identification problem results in a large-scale optimisation problem, we show that, due to a special structure in the linearised Karush–Kuhn–Tucker optimality conditions, the solution can be found very efficiently. The method is applied to a set of test problems relevant for gait analysis. These involve determining the local coordinates of markers placed on the model, segment lengths and joint axes of rotation from both gait and range of motion experiments.

Prediction of ground reaction forces and moments during sports-related movements

“…During the experiment, the subjects performed multiple gait trials of which a single trial for each subject was initially used to determine segment lengths and model marker positions. These parameters were estimated by minimising the least-square difference between model and experimental markers using the method of Andersen et al [35]. For each subject, the segment lengths and marker positions obtained from the gait trial were subsequently saved and used for the analysis of all other trials.…”

“…Model scaling and kinematic analysis were performed applying the optimisation methods of Andersen et al [10,35]. During the experiment, the subjects performed multiple gait trials of which a single trial for each subject was initially used to determine segment lengths and model marker positions.…”

Scaling of musculoskeletal models from static and dynamic trials

“…The model was scaled based on the gait trial itself. The segment length (see Figure 2(B)) as well as local coordinates of markers not placed on anatomical landmarks were identified in an optimisation routine (Andersen et al 2010). A more detailed description of the scaling method and the optimised parameters is supplied as supplementary material.…”

“…In inverse dynamics-based (Damsgaard et al 2006) musculoskeletal models, the modeller typically has to manually scale each body segment and in some approaches also place the skin markers on the model in the locations corresponding to the placement on the subject during a motion capture experiment. Alternatively, optimisation can be used to compute the segment lengths and skin marker locations for a subset of the markers (Andersen et al 2010), but the remaining markers have to be placed manually.…”

A Trap Motion in Validating Muscle Activity Prediction from Musculoskeletal model using EMG

“…Anthropometric data such as body weight, body height, pelvis, thigh, shanks and foot length were imported from the subject measurement. The default scaling algorithm which is based on mass-fat scaling algorithm was applied in the model [1]. In GLEM, the knee was modeled as a hinge between femur and tibia bone.…”