Questions: Leg Muscle Force Estimation

Part I: Leg Muscle Force Estimation in Swing

A.  Explore the model and the Forward Dynamics Tool

1. From the IK results, plot hip flexion, knee angle, and ankle angle as functions of time.
2. From the FWD_locked results, plot hip flexion, knee angle, and ankle angle as functions of time.
3. Discuss the difference in joint trajectories between the inverse kinematic and forward dynamic simulations.
4. Examine the pelvis_tilt, pelvis_tx, and pelvis_ty coordinates of the IK solution, and discuss the effect of locking the coordinates in the forward simulation.

B.  Simulate swing phase with manually selected excitations

1. To match the hip flexion, knee angle, and ankle angle coordinates of the inverse kinematic solution, which three muscles did you excite during the forward dynamic simulation?
2. Continue adjusting your muscle excitations until you are satisfied with the tracking. Plot the excitations of your three muscles as functions of time.
3. Plot the hip flexion, knee angle, and ankle angle coordinates for the IK solution and your forward solution together.
4. Discuss your observations regarding the challenge of tracking trajectories with a forward dynamic simulation.

C. Simulate swing phase with activations from Static Optimization (SO)

1. Plot the muscle activation patterns from the Static Optimization results folder, then plot the activations from the Forward simulation with your best set of controls (from Part B). How do the activations compare?
2. Plot the hip flexion, knee angle, and ankle angle coordinates from the Forward simulation that used controls found by Static Optimization, along with the coordinates from the IK solution. Why does the Forward simulation using the Static Optimization controls diverge from the IK solution?

D. Simulate swing phase with excitations from Computed Muscle Control (CMC)

1. Plot the hip flexion, knee angle, and ankle angle coordinates from the CMC solution along with the coordinates from the IK solution.
2. Plot the hip flexion, knee angle, and ankle angle coordinates from the Forward simulation that used controls found by CMC, along with the coordinates from the IK solution. Are the simulated trajectories better or worse than the Static Optimization results? What is responsible for the difference?

Part II: Leg Muscle Force Estimation in Stance

A. Find a set of generalized forces that produce the stance motion using Inverse Dynamics

1. Plot the applied forces and torques that act on the pelvis (i.e., pelvis_tilt_moment, pelvis_tx_force, and pelvis_ty_force) as functions of time. The data is in the ID/inverse_dynamics.sto file.
2. What does the pelvis_ty_force curve tell you about the force applied during stance?
3. Plot the experimentally measured vertical ground reaction forces (ground_force_vy [right leg] and 1_ground_force_vy [left leg]) from the leg69_stance_grf.mot file and the pelvis_ty_force as functions of time.
4. How do the ground forces compare to the pelvis_ty_force?
5. Given that the model consists of only a single leg and pelvis, in what time range is it reasonable to use this model with the given kinematics and measured forces?

B. Model refinement through residual reduction

1. Why does the pelvis translate substantially in the y-direction (pelvis_ty coordinate) during the simulation?
2. Plot the RRA residual actuator forces (i.e., MZ, FX, and FY) from leg6dof9musc_controls.sto.
3. Open the Messages pane and locate the recommended overall mass adjustment from the last run of RRA (e.g., "pelvis: orig mass = 10.7538, new mass = xxxxx"). Note that the units are in kilograms (kg). What is the recommended mass adjustment? Why would the mass adjustment be so large?

C. Finalize the model through iteration of RRA

1. For your final RRA iteration, plot the tracking error values from the RRA/leg6dof9musc_pErr.sto file.
2. For your final RRA iteration, list the weights of all coordinates in tracking tasks.
3. For your final RRA iteration, which coordinate has the maximum tracking error and what is the maximum tracking error value?

D. Forward simulation of stance with CMC

1. Plot the muscle activation patterns from the states file in the CMC results directory. (Tip: When selecting quantities to plot, use the "Filter By Pattern" to help your search. For example, typing "act" will list only the activation signals.)
2. Are the simulated activations for the vastus intermedius (vas_int_r.activation) and gastrocnemius medialis (med_gas_r.activation) close to what you would expect?
3. Plot the CMC residual joint moments (hip_flexion_r_moment, knee_angle_r_moment, and ankle_angle_r_moment) as functions of time from the *Actuation_force.sto file.
4. Are the moments substantial? Does the plot give you confidence in the predicted muscle forces?
5. Plot the CMC residual actuator forces (i.e., MZ, FX, and FY) as functions of time.

6. How large are the residual joint moments and residual actuator forces predicted by CMC as a percentage of the body weight of the model? Note: Residuals less than 2% of body weight are considered acceptable.

Reflection

Provide some suggestions for future offerings of the course:

• Were any elements of this lab confusing? How could they be improved?
• What resources did you use? Was anything missing or unclear in the OpenSim documentation?
• What was the best part of this lab? What would you add to improve the lab?