Figure 4.

RSM approach for determining the optimal objective function weightings to allow simulation predicted postures to match empirical motion capture data (Davidson et al., 2022). Adapted from Davidson et al., (2021), motion capture data was mapped onto an avatar via anatomical hardpoints. The global position of the avatar’s hands and feet were then used as constraints for the predicted postures. Objective functions weightings were iteratively simulated at a set resolution and each predicted posture was compared to the origin motion capture driven posture to determine an error. All errors for a given posture and prediction set produced a surface which was fit with a regression function and then optimized to find the best objective function weighting.