Inverse kinematics is the problem of posing an articulated figure to obtain a wanted goal,
without regarding inertia and forces. Joint limits are modeled as bounds on individual degrees of freedom,
leading to a box-constrained optimization problem. We present A projected Non-linear Conjugate
Gradient optimization method suitable for box-constrained optimization problems for inverse kinematics.
We show application on inverse kinematics positioning of a human figure. Performance is measured
and compared to a traditional Jacobian Transpose method. Visual quality of the developed method is
evaluated.