Abstract— In this article the principles of human locomotion are revisited and reviewed. This has been done in the framework of two European projects, where the elicitation of these mechanisms inform, on the one hand, the design of artificial bipedal walkers (H2R), and on the other hand the design of lower limb exoskeletons (BETTER) for rehabilitation of gait in post-stroke patients. Passive dynamics emerging from the morphology of the human musculoskeletal system, reflexes as stabilization mechanisms, modular control of movement as well as supra-spinal control of gait are reviewed to get insight on how these mechanisms can be used to explain human locomotion. I. INTRODUCTION The human musculoskeletal and neural-motor system is highly optimized for efficient locomotion. Efficiency, stability and voluntary modulation of human gait are a result of a combination of features spanning the human musculoskeletal, sensorimotor and neural systems. Salient aspects of these systems include (1) the functional morphology, (2) the synergistic coordination of motor activity, (3) the phase dependent modulation of muscle activity and (4) cognitive skills. On the one hand, the functional morphology is highly optimized for efficient biped locomotion [1] as it allows exploiting the inherent dynamics to reduce energy consumption and control effort, and result in natural looking motions. Also a contribution of this functional morphology is the capability of self-stabilization, since the elastic properties of muscles and tendons increase stability without active control. On the other hand, the synergistic feed-forward motor patterns, be it activated at kinetic or kinematic events or due to learned timing, create coordinated synergies of movement. Whilst feedback control occurs at various levels of complexity regarding the extension of perception and deployment of muscle action, phase-dependent modulation is a function of the current task or phase of motion, and as a result reflex action can be modulated, reinforced, or suppressed. Eventually, cognition plays a crucial role in learning and predicting the sensory consequences of actions, helping to deal with feedback time delays and allowing for planning the appropriate compensative actions (active and passive). *Research has been supported by The Spanish CONSOLIDER Program through Grant CSD2009-00067 and by the European Commission FP7 through grants 247935 (BETTER) and 600698 (H2R). J.L. Pons (corresponding author, phone: +34918711900; fax: +34918717050; e-mail: jose.pons@csic.es), J.C. Moreno and D. Torricelli are with the Bioengineering Group, Spanish National Research Council, CSIC, Arganda del Rey, Madrid, Spain. J.S. Taylor is with the Hospital de Parapléjicos de Toledo, Toledo, Spain. This paper will address and review the various motor principles leading to efficient locomotion in humans. In so doing, we will analyze the functional morphology in section II, the contribution of reflex mechanisms to stabilize locomotion in section III, the orchestration and synergistic coordination of motor patterns in section II and eventually the supra-spinal control of gait in humans in section IV. II. FUNCTIONAL MORPHOLOGY OF THE HUMAN MUSCULOSKELETAL SYSTEM Research on the biomechanics of human locomotion provides valuable insights into basic principles for motor control. Human legged locomotion is so efficient partially because it does not power movements with independent motor actions at each joint. Muscles often span multiple joints, which results in energy-saving and power transfers when a movement simultaneously requires negative power at one joint and positive power at another joint [2]. This allows making effective use of passive elastic properties to generate part of the required force or power without metabolic cost, especially when muscle-tendon units span multiple joints [3]. Furthermore, multiple muscles spanning a joint allow efficient modulation of joint stiffness during dynamic movements, making fast adaptation to uneven surfaces and terrains possible. Passive dynamic walking introduced, as a model, in under-actuated bipedal walkers and robots shows emerging, natural-looking walking gait with remarkable similarities to human walking. In this regards, passive dynamic walking, first introduced by McGeer [4], exploits the mechanical potential energy gained while walking down a slope, showing a stable gait without any or limited control or actuation. Limit Cycle (LC) walking machines represent a step forward in this direction. They combine the exploitation of passive dynamics with minimal feed-forward actuation in order to replenish energy losses and to increase stability. Examples of LC walking prototypes have been developed by TU Delft [5], Cornell University and MIT [6]. The Cornell and Delft bipeds demonstrate that basic walking can be accomplished with extremely simple control and very low energy consumption. However, due to absence of feedback control, passive dynamics bipeds cannot react to disturbances or external forces, even though human-like gait is achieved, walking is unstable as all other stabilization mechanisms found in humans are still lacking. III. ROLE OF REFLEX FUNCTION FOR STABILIZATION OF LOCOMOTION Bipedal plantigrade walking in humans is unique compared to digitigrade locomotion in animals, in that the stance phase is made by ground contact first by the heel, then the foot sole and finally the toes. In addition support is Principles of human locomotion: a review* J.L. Pons, Member, IEEE, J.C. Moreno, D. Torricelli and J.S. Taylor