Despite decades of study, the mechanisms that determine human locomotion are still unknown, available models and motor control theories can only partially capture the phenomenon. This is probably the principal cause of the reduced efficacy of lower limbs rehabilitation therapies. Recently, it has been proposed that human locomotion may be planned in the task-space by taking advantage of the gravitational pull acting on the Centre of Mass (CoM) that we have used to formulate a task-space planner for straight locomotion at a constant speed. The proposed model represents the CoM transversal trajectory as simple harmonic oscillator moving forward at a constant speed. On the other hand, the vertical trajectory of the CoM is controlled through the ankle strategies. Our solution is composed of closed-form equations which can plan human-like trajectories for both the CoM and the foot swing. The model output can be seen as the optimal trajectory determined based on the average behaviour of 12 healthy subjects walking at three self-selected speeds. Furthermore, the planner formulation is compatible with an extended formulation of the Passive Motion Paradigm which enables us to design a hierarchical architecture of semi-autonomous controllers. The final architecture can also describe the motor primitives as a particular case of dynamic primitives, shows strong parallels with the nervous system organization, and is compatible with the optimal feedback controller theory.