Control of human gait stability through foot placement

Abstract

During human walking, the centre of mass (CoM) is outside the base of support for most of the time, which poses a challenge to stabilizing the gait pattern. Nevertheless, most of us are able to walk without substantial problems. In this review, we aim to provide an integrative overview of how humans cope with an underactuated gait pattern. A central idea that emerges from the literature is that foot placement is crucial in maintaining a stable gait pattern. In this review, we explore this idea; we first describe mechanical models and concepts that have been used to predict how foot placement can be used to control gait stability. These concepts, such as for instance the extrapolated CoM concept, the foot placement estimator concept and the capture point concept, provide explicit predictions on where to place the foot relative to the body at each step, such that gait is stabilized. Next, we describe empirical findings on foot placement during human gait in unperturbed and perturbed conditions. We conclude that humans show behaviour that is largely in accordance with the aforementioned concepts, with foot placement being actively coordinated to body CoM kinematics during the preceding step. In this section, we also address the requirements for such control in terms of the sensory information and the motor strategies that can implement such control, as well as the parts of the central nervous system that may be involved. We show that visual, vestibular and proprioceptive information contribute to estimation of the state of the CoM. Foot placement is adjusted to variations in CoM state mainly by modulation of hip abductor muscle activity during the swing phase of gait, and this process appears to be under spinal and supraspinal, including cortical, control. We conclude with a description of how control of foot placement can be impaired in humans, using ageing as a primary example and with some reference to pathology, and we address alternative strategies available to stabilize gait, which include modulation of ankle moments in the stance leg and changes in body angular momentum, such as rapid trunk tilts. Finally, for future research, we believe that especially the integration of consideration of environmental constraints on foot placement with balance control deserves attention.

Keywords: balance; bipedal walking; foot placement; gait stability.

Figure 1.

Example of mediolateral CoM motion and foot placements during normal gait. Just after midstance (coloured dots in CoM trace), the CoM starts moving to what is to become the new stance foot. Around the peak CoM velocity (not shown), the next foot placement happens, and the CoM motion starts to be redirected to the next foot. (Online version in colour.)

Figure 2.

(a) The inverted pendulum model as used in many studies. In this model, the pendulum is assumed to rotate around the ankle joint, which is based in the foot segment, which has a certain BoS. All of the models we describe here take into account both the position (CoM) and velocity (V_CoM) of the CoM. The FPE also takes into account the angular momentum around the CoM (I*ω). Furthermore, most methods further simplify this model by assuming that changes in the ankle angle (φ), only change the horizontal position (and velocity) of the CoM, not the vertical. Note that we draw here an AP schematic, but the same schematic holds for ML. (b) Finally, only the FPE method takes into account that when the BoS is shifted to regain stability, this coincides with a collision (impact impulse), which leads to the fact that the velocity after the collision is lower than the velocity before the collision , and hence, when no energy would be added to the system, the XCoM concept (and other similar concepts that do not model the impact) may predict that stability may be achieved, this may not be so.

Figure 3.

Example of a lateral stabilization set-up as used in several studies. The subject is placed in a frame, which is attached with elastic bands to the outside world. These cords are either very long or are attached to sliding rails (such as shown here), such that they do not interfere with the AP motion of the subject.