This paper applies a robotics-inspired approach to derive a low-dimensional forward-dynamic hybrid model of human walking in the sagittal plane. The low-dimensional model is derived as a subdynamic of a higher-dimensional anthropomorphic hybrid model. The hybrid model is composed of models for single support (SS) and double support (DS), with the transition from SS to DS modeled by a rigid impact to account for the impact at heel-contact. The transition from DS to SS occurs in a continuous manner. Existing gait data are used to specify, via parametrization, the low-dimensional model that is developed. The primary result is a one-degree-of-freedom model that is an exact subdynamic of the higher-dimensional anthropomorphic model and describes the dynamics of walking. The stability properties of the model are evaluated using the method of Poincaré. The low-dimensional model is validated using the measured human gait data. The validation demonstrates the observed stability of the measured gait.