Predicting the Biomechanics of Walking - Human Motion Simulation: Predictive Dynamics - page 182

7.13.5 Walking on uneven terrains
While only simple forward walking was presented here, the method has been
extended to walking on uneven terrain (Hariri, 2011). The major change in the
formulation is in the ZMP approach that was originally developed for forward
walking on a flat surface. Walking on uneven terrains includes stair climbing and
many other implementations.
7.13.6 Asymmetric walking
Asymmetric walking can be simulated by formulating the problem for a full stride
as demonstrated by Xiang et al. (2009a,b, 2010a,b,c) .
7.13.7 Walking on different terrain types
Foot contact with the ground, including friction conditions and various ground
types such as mud or ice, is important in many fields. The interaction between the
foot and soft inelastic material such as gravel is an area of future research.
7.14 Concluding remarks
This chapter is dedicated to the detailed treatment of formulating a predictive
dynamic walk cycle. The main objective is to demonstrate how gait prediction is
formulated and how it can be used to study cause and effect. We summarize the
main advantages of using predictive dynamics as a method for understanding
motion and the effect of physics in modeling and simulating human motion.
a. Walking with various forms, loads, anthropometries, and ranges of motion can
be simulated.
b. Integration of the equations of motion is not carried out. As a result, the
prediction of a walk cycle is faster and fundamentally of higher fidelity
because of the optimization algorithm.
c. The cause and effect can be obtained by varying the input to the model.
d. A high number of DOFs for the human body (55 DOF were used) can provide
an adequate, natural, and realistic motion.
e. Global DOFs and virtual joints can be used to provide for the global
transformation motion of the digital human with respect to a fixed reference
frame.
f. The objective function used as a characterization of the joint torque square
provides a driving function for the motion, to enable a natural human-like
realistic behavior.
g. Ground reaction forces using the predictive dynamics formulation can be
recovered.