This paper presents a method to determine outputs associated with human walking data that can be used to design controllers that achieve human-like robotic walking. We consider a collection of human outputs, i.e., functions of the kinematics computed from experimental human data, that satisfy criteria necessary for human-inspired bipedal robot control construction. These human outputs are described in a form amendable to controller design through a special class of time based functions - termed canonical walking functions. An optimization problem is presented to determine the parameters of this controller that yields the best fit to the human data that simultaneously produces stable robotic walking. The optimal value of the cost function is used as a metric to determine which human outputs result in the most “human-like” robotic walking. The human-like nature of the resulting robotic walking is verified through simulation.