Unbalance detection to avoid falls with the use of a smart walker

Abstract

Smart walkers aimed at providing better support than conventional walker. These devices have synchroneous movements with users but they have to act differently in case of unbalanced gait. The available discriminatory factors used in smart walkers are generally based only on position data. These static data cannot be representative of the dynamical process of falling. This paper proposes three methods that could be used to detect unbalanced gait. They are based on different kinematic data: forward velocity, angular velocity around transverse axis and eXtrapolated Center Of Mass (XCOM) with stability margins. These methods are evaluated and compared experimentally on 4 young healthy subjects experiencing unexpected unbalanced gait when using a robotic walker.