Human-in-the-loop optimization of rocker shoe to reduce plantar pressure and collision work simultaneously

Abstract

Background

Rocker shoes can be used to reduce foot pressure and adjust lower limb kinetics for various patient population, such as people with diabetic peripheral neuropathy. Selecting adequate properties of the rocker sole is of great importance for its efficacy. This study investigated the capability of human-in-the-loop optimization (HILO) to individually optimize apex position and angle of rocker shoe to reduce peak pressure and collision work simultaneously.

Methods

Peak pressure, kinetic, and kinematic data were recorded from 10 healthy participants while walking at preferred speed wearing rocker shoes with adjustable apex position and angle. An evolutionary algorithm was used to find optimal apex parameters to reduce both peak pressure in medial forefoot and collision work. The optimized shoe (HILO shoe) was compared with generic optimal rocker settings (Chapman settings) and normal shoe.

Findings

Compared to normal shoe, the HILO shoe had lower plantar pressure (p_HILO = 0.007; p_Chapman = 0.044) and Chapman shoe showed higher collision work (p_HILO = 0.025; p_Chapman = 0.014). Both HILO and Chapman shoe had smaller push-off work than normal shoe (p_HILO = 0.001; p_Chapman < 0.001) with the Chapman shoe exhibited earlier push-off onset (p_HILO = 0.257; p_Chapman = 0.016).

Interpretation

The Human-in-the-loop optimization approach resulted in individualized apex settings which performed on average similar to Chapman settings but, were superior in selected cases. In these cases, medial forefoot could be further offloaded with apex angles larger than generic settings. The larger apex angle might increase the external ankle moment arm and push-off work. However, there is limited room for improvement on collision work compared to generic settings.