Skip to main content

Model Requirements

For full or lower body models, we recommend using the RCNL model available with the pipeline examples, based on the Rajagopal 2016 model. Before using the pipeline, the model should be properly scaled.

Joint Model Personalization​

  • For lower body models, use the RCNL model, which has modified knee and ankle joint parameters to improve modeling kinematics with Joint Model Personalization.
  • Marker names in the model must be consistent with marker names in the provided .trc files.
  • It is recommended to have at least three tracked markers on each body segment adjacent to a joint being optimized.

Muscle-Tendon Personalization​

  • The following four muscle group types must be defined in the model:
    • Measured EMG Groups: For every EMG data column that will be used, there must be a muscle group in the model sharing the name of the data column. This group should contain every muscle the EMG channel describes.
    • Missing EMG Groups: If synergy extrapolation is used, muscles not included in measured EMG channel groups must be placed in a missing EMG channel group. Muscles that will share estimated EMG data should be in the same group.
    • Activation Groups: Muscles that should maintain similar activations to each other for the given data should be grouped together. The cost function will penalize deviations in activation within each group.
    • Normalized Fiber Length Groups: Muscles that should maintain simiar normalized fiber lengths to each other for the given data should be grouped together. Similar to the activation groups, the cost function will penalize deviations in normalized fiber length within each group.
  • For Muscle-Tendon Personalization, every muscle being optimized must be included in either a measured EMG channel group or a missing EMG channel group. Muscles not included in activation groups or normalized fiber length groups will not be used in those grouped cost terms.
  • Muscle strength (maximum isometric force) should be set/scaled appropriately for the model being used before starting Muscle-Tendon Personalization.
    • While this property may be adjusted by Muscle-Tendon Length Initialization, these changes are only intended to maintain consistency with calibrated optimal fiber and tendon slack lengths.

Neural Control Personalization​

  • The following three muscle group types must be defined in the model:
    • Synergy Set Groups: In Neural Control Personalization (NCP), multiple synergy sets may be used. For each synergy set, a group sharing the name of the synergy set defined in the NCP settings must be created, and it must contain every muscle the synergy set describes.
    • Activation Groups: Muscles that should maintain similar activations to each other for the given data should be grouped together. The cost function will penalize deviations in activation within each group.
    • Normalized Fiber Length Groups: Muscles that should maintain simiar normalized fiber lengths to each other for the given data should be grouped together. Similar to the activation groups, the cost function will penalize deviations in normalized fiber length within each group.
  • For NCP, every included muscle must be placed in one of the synergy set groups.
  • Muscle strength (maximum isometric force) should be set/scaled appropriately for the model being used before starting NCP.
    • While this property may be adjusted by Muscle-Tendon Length Initialization, which is also used by NCP, these changes are only intended to maintain consistency with calibrated optimal fiber and tendon slack lengths.

Ground Contact Personalization​

  • Any foot that will have a ground contact model defined must be a two-segment foot model. Each foot must contain a single toes body connected to a single hindfoot body by a pin joint.
  • The following five markers must be included on each foot:
    • Toe marker: at the distal end of the toes body, attached to the toes body
    • Medial marker: on the medial side of the foot, near the toe joint, attached to the hindfoot body
    • Lateral marker: on the lateral side of the foot, near the toe joint, attached to the hindfoot body
    • Heel marker: behind the heel, attached to the hindfoot body
    • Midfoot superior marker: near the middle of the foot, superior to the bodies of the foot in a standing pose, attached to the hindfoot body
  • The five foot markers may or may not be the same markers used in Joint Model Personalization. If the markers were not used for motion capture, their positions can be estimated based on our example models and placed on the model manually if needed. Note that ground reaction moments will be calculated and reported about the midfoot superior marker projected onto the floor.