Update: September 7, 2011
Corrected a bug in the knee_angle_r and knee_angle_l coordinate ranges 
that was introduced in the 2-leg, OpenSim versions of this model.  The 
range was increased from 0-100 degrees to 0-120 degrees, which 
introduced invalid moment arms at knee angles above 100.  The error has 
been corrected by reducing the range of motion to the original 0-100 
degrees, which was validated against experimental measurements of moment 
as described the above publication.
Additionally, wrapping surface display properties have been hidden by 
default.

Update: April 12, 2011
version 2: The patellar joint and corrseponding coupler constraints were 
altered to work around a problem with coordinate coupler constraints 
that occurs during scaling  in OpenSim.

Translational coordinate coupler constraints were scaled incorrectly in 
the OpenSim 2.2 Scale tool.  This affected constraints that were used to 
define patellar motion as a function of knee_angle_l and knee_angle_r.  
This issue is worked around here by replacing the three coordinates 
defining rotational and translational motion in the left and right 
patella bodies with a single rotational coordinate in each, 
knee_angle_beta_*, and coupling the new coordinates to equal 
knee_angle_*.

This model includes the following additions and changes relative to 
the single-leg version 2:

Left leg with same architecture and mirrored geometry as right leg.

6 degrees of freedom for the pelvis segment (translational x,y, and z)
rotational tilt, list, and rotation.

Torso segment with composite head-arms-torso (HAT) mass.

3 degree of freedom lumbar joint (extension, bending, and rotation),
Torso muscles with architecture as defined in the OpenSim default 
model gait_2392.

Activation dynamics parameters to produce same activation/deactivation
timing for all muscles:
Schutte1993Muscle (activation1 = 7.6, activation2 = 2.5)
Thelen2003Muscle (activation_time_constant = 0.01,
deactivation_time_constant = 0.04)

Markers set placed according the convention used at the University of
Delaware Neuromuscular Biomechanics Lab.

Two strength levels: 
Default - the <max_isometric_force> values are the same as reported in 
the paper and used in the single leg version.
50% Stronger - All lower limb muscles (i.e. not back muscles) have 
been strengthened by increasing <max_isometric_force> 50%.  This model 
is strong enough to produce a CMC simulation of walking for an average 
sized, healthy, adult male.