From left to right: Peter Cavanagh, Jill Startzell, Nori Okita, David Lemmon.

A project by the Center for Locomotion Studies (CELOS), Penn State University, University Park PA 16802

Figure 1: Conceptual diagram of stair laboratory at CELOS

Locomotion on stairs is among the most challenging and hazardous activities of daily living that elderly individuals encounter. Many deaths and injuries occur among the elderly on stairs and it is not well known that accidents during descent outnumber those during ascent by more than three to one.  Fear of stairs can severely limit the mobility and socialization of elders. Researchers at Penn State University are studying stair descent under a grant from the National Institute on Aging to understand more about the mechanisms of stair accidents and to devise ways to make stair locomotion safer for the elderly.

Figure 2: Vicon cameras mounted from a Unipod on a Bogen arm.

Much work in the literature on stair walking has examined negotiation of a flight consisting of only a few stairs.  However, research by Templer (1992) suggested that gait in the different phases of stair descent depends on quite different sensory cues, and to examine this further, the Penn State group decided to build a stair laboratory which is shown schematically in Figure 1.  The central feature is a flight of seven stairs with adjustable rise and run (height and depth).  Kistler force platforms are built into the tread of steps 2 and 4, and a larger force platform is in the ground on the bottom landing.  A Vicon 370 motion analysis system is used with six 60Hz progressive cameras in high spatial resolution mode to collect kinematic information from marker clusters mounted from the head to the toe of a walking subject.  Six additional standard 60Hz cameras are also available and can be switched into the Vicon 370 system to provide collection of kinematic data during overground walking in the adjacent gait lab.   Graduate student Justin Guerin has devised novel Unipods on which the cameras in the gait area are mounted (Figure 2) using Bogen arms for easy adjustment.  Cameras in the stair area are either mounted overhead from Unistrut beams or from floor mounted Unipods.

Figure 3: A small clearance of the shoe sole with respect to the stair edge during swing.

Unexpected contact between the foot and the stair can result in a stumble or fall, potentially leading to serious injury or death. Thus the measurement of clearance may improve the understanding of how falls occur and what functional adaptations are made to make stair negotiation strategies in challenging conditions. Foot clearance during stair walking can be defined as the closest distance between any point on the outsole of the shoe and the staircase during the swing phase.  Depending on the gait style of an individual, minimal clearance can occur relative to the stair edge or tread, or even on the riser for those stair users who choose to detect the geometry of the stair explicitly (by Ôfeeling’ the stair dimensions).  In order to accurately measure clearance in three dimensions, a series of calibrations are performed to define the global environment, the critical stair vertices, the local shoe reference frame, and a mesh of coordinates which describe the outsole surface relative to fixed shoe markers (Figure 3).

With this information, three dimensional marker data can be transformed into meaningful positional data describing the orientation of the shoe surface relative to the staircase.  

Figure 4: Static calibration configuration.

However, this application demands extremely high accuracy in a rather large and irregularly shaped field of view (3.2 m x 3m x 1.2m).  Previous work by this group (Simoneau et al. 1991) has shown that clearance can be as small as 3 mm under certain circumstances (Figure 3).  The lab uses DynaCal, and their first attempt to calibrate the space used a conventional gait lab L-frame mounted on its side to ensure marker visibility.   Research at Oxford Metrics showed that changes in orientation of the L-Frame as small as 3 millirads (approximately 0.2 degrees) in this configuration could result in errors of up to 9 mm at a distance of 3 metres from the origin of the global reference frame.  Thus, the new calibration configuration shown in Figure 4 was developed.  Three reflective spheres 10cm in diameter are suspended to form the tripleton needed by the DynaCal static calibration procedure.  The singleton is a fourth sphere placed approximately 2.5 metres from the hanging set.  Static points are defined in the field using a wand (Figure 5) similar to that proposed by Cappozzo et al. (1995).  Using this approach, static accuracy of better than 2mm throughout the field has been demonstrated (a resolution of 1 in 1600).

Figure 5: CELOS faculty member Stephen Piazza uses a wand to determine the location of calibration points on a test linkage simulating the knee joint.

Most post-processing of data is con-ducted using 4 x 4 matrix algebra coded in Matlab routines (see Hyperlink http://www.celos.psu.edu/kinematics).  For example, the clearance algorithm (Startzell and Cavanagh 1999) shown schematically in Figure 6 examines the clearance of several hundred virtual points on the shoe outsole in relation to the stair at every frame.  The 3D wire frame stick figure of the three lower extremity segments shown in Figure 7 is simply a reconstruction from an array of anatomical-to-global transformations for each segment at each frame.

Much of the experimental work is concerned with the analysis of biomechanical responses to sensory challenge.  For example, early work with young subjects has shown trends toward safer strategies (such as increased clearance) under challenging conditions induced by high speed and low light-ing (Startzell, 1998). This group of researchers has now been joined by doctoral student Kate Christina, who together with consultant architect John Templer, geriatrician Stephanie Studenski, visual psychologist Fred Owens, and statistician Steve Arnold, will contribute to the exploration of safety during stair descent in the elderly over the next several years.

References:

1.       Cappozzo A, Catani F, Della Croce U, Leardini A.  (1995)       Position and orientation of bones during movement.  Anatomical frame definition and determination.  Clin. Biomech: 4: 171-8.

2.       Startzell JK and Cavanagh PR. (1999) A three dimensional approach to the calculation of foot clearance during locomotion. Human Movement Science: In Press.

3.       Startzell JK. (1998) Foot clearance and placement during stair descent.  The effect of speed and illuminance.  Master Thesis, Penn State University.

4.       Templer J. (1992) The Staircase: Studies of Hazards, Falls, and Safer Design. Cambridge, Mass. MIT Press.

5.       Simoneau GG, Cavanagh PR, Ulbrecht JS, Leibowitz HW, Tyrrell RA. (1991) The influence of visual factors on fall-related kinematic variables during stair descent by older women. J Gerontol;46(6): M188-M195.