Fig 1.
Experimental setup and schematic of the subjects initial position.
Measuring setup including the KUKA KR270 2700 ultra with the AMTI OR6 Series Force Plate mounted to the end effector, the seat for the participants and the monitor for visual feedback. The schematic in the upper left corner shows the intended sagittal angles in the hip, knee, and ankle joint for the subject’s initial position with the backrest set to 70° uprise.
Fig 2.
Experimental marker distribution throughout the lower limb showing landmark and cluster markers in initial pose with force (green arrow) applied (left) and in neutral standing pose (right).
Fig 3.
Scaling and individualization processes and resulting model.
(A) The schematics describe the scaling processes for the Osim FUN 1 and Osim FUN 2 models. The processes are, respectively, embedded in the overall workflow as depicted in S1 Fig of the supplementary material. Left (Osim FUN 1): the iterative segment by segment (from proximal to distal) approach based on [44]. First, the generic OpenSim model contains only markers of the proximal segment of the first joint (hip) to be individualized. This segment was scaled and the markers were adjusted (using OpenSim scaling tool). Second, that joint’s calculated FARs were implemented (using MATLAB) and, last, the markers of that joint’s distal segment were added (using OpenSim scaling tool). The upcoming iterations of the process were performed with the next joints (knee and ankle). Right (Osim FUN 2): the all in one approach. The generic model contains all markers for all segments. The complete model was scaled and all markers were adjusted before implementing the FARs for both the proximal and distal segments of each joint. (B) OpenSim’s scaled musculoskeletal unilateral lower extremity model including the FARs and local coordinate systems (LCS) for the knee and ankle joint and the femur and tibia segment showing the axes of the functional joint coordinate systems (e.g. ) and the axes of the respective segment coordinate systems (e.g.
) for the joints proximal segment.
Table 1.
Included markers (indicated with ) in the marker sets S1 to S8 used in the sensitivity analysis. Markers were placed at anatomical landmarks and in clusters on the pelvis and right extremity segments. The standard marker set used for the model evaluation regarding the EAM was the full marker set S8.
Fig 4.
External knee adduction moment.
Boxplots representing the (A) EAM values calculated for the three tested foot positions for the static trials, (B) the average and (C) the maximum EAM of the dynamic legpress trials with the REF model (red), the Osim CON model (purple), the Osim FUN 2 model (green), and the Osim FUN 1 model (blue). Black horizontal line in the box represents the median, the whiskers are defined by the (interquartile range). Black dots represent individual data points. Significant differences are stated in the text and Table 2.
Table 2.
The table presents the results of the ANOVA (last column) with the model as the main factor, and the corresponding post hoc analysis. Values are shown for the EAM in percentage of body weight times height () during static trials (top section), the mean EAM during dynamic trials (middle section), and the maximum EAM during dynamic trials (bottom section).
Table 3.
RMSE of the difference, mean absolute difference, and maximum absolute difference of the EAM between the Osim FUN 2 model using the marker sets S1 to S8 and the REF model calculated for the dynamic trials.
Fig 5.
Osim RT vs Osim FUN 1 joint moments.
The EAM (A), the ankle joint dorsiflexion moment (B), the knee joint flexion moment (C), and the hip joint flexion moment (D) as functions of the normalized leg extension movement calculated with the Osim FUN 1 (blue) and the Osim RT model (dark blue). The lines represent the respective average moments, and the accordingly colored areas represent their standard deviation. The visualization of the Osim model represent the start and end position at and
of the leg extension movement, respectively.