Analysis of gait event detection algorithms applied to movement data collected on a sloped treadmill

In gait analysis, it is important to report kinematics and kinetic results delimited by specific gait cycle events. In particular, foot-contact and foot-off events are used to define the phases of each stride, and to segment a stream of movement data into discrete, meaningful sections that can be analyzed and compared. In a laboratory environment where adequate force platforms are available, acquiring these events does not pose a problem. However, motion analysis data is also collected outside of this ideal environment and the same need for accurate identification of these events still applies. This paper analyzes several algorithms for identifying gait events from movement data and applies them to treadmill walking. Treadmills pose a challenge to identification algorithms due to the lack of force plate information and the constant motion of the walking surface. A multi-feature algorithm that successfully identifies gait events from treadmill data is developed and tested.