Original Article J. Phys. Ther. Sci.

14: 33–39, 2002

Relationship between the Required
Coefficient of Friction and Gait Initiation in
Young Adults on a Low Friction Floor

TADAYOSHI ASAKA, RPT, PHD1), HIROSHI SAITO, RPT, MS1), NAOKI YOSHIDA, OTR, MS2),
DAISUKE URAKAMI, RPT, MA1), KOJI KAMADA, RPT1), JUNKO FUKUSHIMA, MD, PHD1)
1)
Department of Physical Therapy, College of Medical Technology, Hokkaido University:
N12 W5 Kita-Ku, Sapporo City, Hokkaido 060-0812, Japan. TEL & FAX +81 11-706-3391
2)Department of Occupational Therapy, College of Medical Technology, Hokkaido University

Abstract. The objective of this study was to clarify how movement is controlled at gait initiation in order
to avoid increasing required coefficient of friction (RCOF) on a low friction floor. We aimed at examining
the relationship between RCOF during push off for the second step and movement of the first step. Twelve
healthy young adults underwent measurement of ground reaction forces during push off for the second step
using a force plate. Movement of the first step was measured with a motion analysis system. Upon
initiation of the gait on the low friction floor, frequency and displacement of slips were greatest during push
off for the second step. As the step length for the first step decreased, there was a significant decrease in
the maximum value of RCOF during push off for the second step. These findings suggest that the step
length for the first step is reduced in order to avoid increasing RCOF during push off for the second step.
Key words: Gait initiation, Required coefficient of friction (RCOF), Low friction floor.

(This article was submitted Nov. 21, 2001, and was accepted Dec. 25, 2001)

INTRODUCTION friction between the sole and floor (hereafter

referred to as the required coefficient of friction:
The risk of bone fracture is high when the elderly RCOF) is calculated by dividing the horizontal foot
fall, and the prognosis is poor for elderly patients force by the vertical foot force. Several studies
with a bone fracture. Healthcare for the elderly have reported a relationship between slipping and
following bone fracture is recognized as a social RCOF8–10). Lockhart11) reported that, unlike the
problem that has taken on increasing importance young, the elderly are less able to reduce RCOF at
with the “graying” of society1, 2). Slipping and heel strike on a low friction floor even when the
tripping are two major factors associated with falls cadence increases. It has been reported that during
among the elderly3–5). the cyclic movement of walking (steady gait) on a
Moreover, in colder regions where snow and ice low friction floor, stride length is reduced to lower
are part of the climatic condition, the elderly and the both the foot velocity and horizontal foot force in
disabled often feel anxious about slipping on frozen order to lower RCOF11, 12). However, to the best of
surfaces and therefore avoid going outside, which our knowledge, no studies have investigated the
lowers their quality of life6, 7). relationship between the movement parameters and
A person “slips” when the horizontal force RCOF.
applied by the foot exceeds the frictional force When carrying out movement analyses on low
between the foot and the surface. The coefficient of friction floors, many authors have investigated
34 J. Phys. Ther. Sci. Vol. 14, No. 1, 2002

posture reactions in response to a slip at heel strike collected into a personal computer. The force plate
during steady gait when transitioning from a high was used to trace center of pressure (COP) while
friction floor to a low friction floor13–16). Ashley et standing and ground reaction forces during push-off
al. 17) reported that the elderly often fall when of the support leg at a frequency of 1,200 Hz. A
initiating and terminating gait or when changing step-off table (600 × 1500 × 35 mm) was designed
direction, which indicates that studying gait specially to match the height of the force plate and
initiation on a low friction floor is important for was then placed adjacent to the force plate in the
preventing falls among the elderly. Miller and direction of intended stride. The subjects stood on
Verstraete 18) analyzed the gait from postural the force plate and swung the first step on the table.
standing to the fifth step in young adults. He found Slips on the low friction floor and kinematic
that the forward motive force increased to the parameters of the first step were measured using the
greatest extent during push off for the second step motion analysis system, and three-dimensional
(support leg), while the first step (swing leg) acted coordinates were obtained from the light reflective
as the base of push-off for the second step. We, markers using two CCD cameras at a frequency of
therefore, aimed at RCOF during push-off for the 60 Hz. Light reflective markers were placed at
second step in this study. eight anatomical landmarks: vertex, right acromion,
In order to understand movement control in right greater trochanter, right knee fissure, right
elderly subjects on a low friction floor, we lateral malleolus, right heel, head of the right fifth
examined young subjects for initial basic data. It is metatarsal, and the head of the left first metatarsal.
necessary to fully understand how young subjects The axes of the two cameras were set at 45 degrees
control gait movement on a low friction floor before from the normal to the subject’s right side so that
examining that of elderly subjects. every marker placed on the right side of the body
The objective of this study was to clarify how could be detected.
movement is controlled at gait initiation in order to
avoid increasing RCOF on a low friction floor in Procedures
young subjects. We examined the relationship Each subject was asked to wear a black T-shirt
between RCOF during push-off of the support leg and a pair of shorts and to stand on the force plate
and movement of the swing leg. The results of this barefoot with arms crossed across the chest so that
study might be applicable to gait training for the the marker on the right greater trochanter was
prevention of slips at gait initiation on a low friction visible (referring to the study of Tang and
floor. Woollacott13)). While standing, the feet were 8 cm
apart and positioned to match the lines indicating a
METHODS foot angle of 10 degrees. Next, the subject was
instructed to start walking at a comfortable speed
Subjects while looking straight ahead horizontally after
The subjects were 12 adult men who were healthy hearing a beep sound. The first step was the right
undergraduates without any past history of leg. All of the data were collected for three seconds
neurological and musculoskeletal diseases. The after the beep sound.
average age, height, body weight and foot size of The low friction floor was prepared as follows: a
the subjects were 23.5 ± 2.8 years, 168.3 ± 5.9 cm, vinyl sheet with a thickness of 0.085–0.086 mm was
60.4 ± 3.9 kg and 25.3 ± 1.8 cm (mean ± standard placed on the force plate and the step-off table, and
deviation). Informed consent was obtained from all then liquid paraffin was applied evenly on top of the
subjects prior to participation in the study. vinyl sheet. Each subject underwent ten
experiments on the force plate and table (high
Materials friction floor) and then ten more experiments on the
The experimental apparatus consisted of a Kistler low friction floor. Each subject practiced several
force platea (600 × 400 × 35 mm) and a motion times before the experiments under both conditions.
analysis system (Frame-DIAS3D) b (Figure 1). While the subjects were on the low friction floor,
Ground reaction forces and three-dimensional one of the examiners was positioned to the left of
coordinates of the markers were synchronized and the subject to support him in the event of a slip.
 35

Fig. 1. Schematic drawing of the experimental set-up.

Subjects stood on top of the force plate and began
walking toward the step-off table using the right leg.
A light reflective marker was placed on eight
a n a to m ic a l la n d m a rk s . Tw o c a m er a s we r e
positioned on the right side of the subject.

Data analysis
As shown in Fig. 1, the data was recorded while
defining the left-right, forward-backward and up-
down directions as x, y, and z coordinates,
respectively: (1) left-right position of COP (COPx) Fig. 2. Various kinetic and kinematic parameters of gait
and forward-backward position of COP (COPy); (2) initiation. Phase 1 started with a change in COPy and
horizontal component force of the left-right ended with a change in the y coordinate of the right
heel marker. COPy max was defined as the
direction of a ground reaction force (Fx), horizontal maximum COPy in Phase 1. Phase 2 started at the
component force of the forward-backward direction end of Phase 1 and ended when the y coordinate of
of a ground reaction force (Fy), vertical component the right heel marker reached a maximum. Step
force of a ground reaction force (Fz); and (3) three- l e ng t h w a s d e f i n e d a s t h e a m o un t of t h e y
dimensional coordinates of the eight markers (x, y, displacement of the right heel marker in Phase 2, and
step time as the length of time required to complete
z). Data was recorded starting with the step length. Vertical displacement of the lateral
displacement in COP and ending with the loss of malleolus (VDLM) was defined as the maximum
ground reaction forces (total time). vertical movement of the right lateral malleolus
The recorded data was analyzed as follows (Fig. marker. Phase 3 started at the end of Phase 2 and
2). First, in order to divide total time into three ended when the floor reactive force reached zero.
Also, H-TCT was defined as the length of time
phases, start and finish times were determined for
between the heel contact for the first step (right leg)
each phase. This division was required for and the contact of the right toe on the floor in Phase 3.
determining various movement parameters. Phase
1 started with the change in COPy and ended with
the change in the y coordinate of the right heel
marker. Phase 2 started at the end of Phase 1 and the right heel marker. Phase 3 started at the end of
ended when they coordinate of the right heel marker Phase 2 and ended when the ground reaction forces
reached a maximum. However, when a subject reached zero.
slipped at heel strike, the end point of Phase 2 was Next, slip frequency and slip displacement were
determined based on the speed of the z coordinate of calculated. Slip displacement was defined as the
36 J. Phys. Ther. Sci. Vol. 14, No. 1, 2002

amount of horizontal shift (∆y) of the foot in the

sagittal plane on the floor under one of the
following three cases: (1) during push-off for the
first step (right leg), ∆y of the right fifth metatarsal
marker in Phase 2; (2) at the heel contact for the first
step, ∆y of the right heel marker in Phase 3; and (3)
during push-off for the second step (left leg), ∆y of
the left first metatarsal marker in Phase 3.
Finally, kinetic and kinematic parameters at gait
initiation were analyzed. The maximum value of
RCOF (RCOFmax) was calculated based on Fy/Fz Fig. 3. Slip frequency and slip displacement for three
in Phase 3. Next, the following five parameters that phases of gait initiation (mean ± S.D.). Among
push-off for the first step, heel contact for the first
could have influenced RCOF were measured: (1) step and push-off for the second step, the slip
the length of the first step by the swing leg, in other frequency and slip displacement were greatest for
words, a change in the y coordinate of the right heel push-off for the second step (40% and 101.0 ± 51.6
marker in Phase 2 (step length); (2) the length of mm, respectively). Right ordinates indicate the
time required for the swing leg to complete the first distance of slip, while left ordinates indicate the
rate in the occurrence of slip.
step in Phase 2 (step time); (3) the maximum
vertical distance between the foot of the first step
and the floor, in other words, the maximum
displacement of the z coordinate of the right lateral push-off for the first step.
malleolus marker; (4) the length of time from heel RCOFmax and length of time from heel contact
contact to toe contact for the first step; and (5) the to toe contact on the low friction floor were
maximum displacement of the y coordinate of COP calculated, excluding those for slip samples at heel
in Phase 1 (COPy max). Step length and vertical strike for the first step because the slip affected
displacement of the lateral malleolus were these two parameters. Step length, step time,
normalized based on height and COPy-max based vertical displacement of the lateral malleolus and
on foot size. COPy max on the low friction floor were calculated
A paired t test was used to analyze differences in for the non-slip and slip groups together because
the average of these five parameters and RCOFmax slips did not occur during push off for the first step.
between high and low friction floors. Pearson’s The RCOFmax for push-off for the second step on
correlation test was used to analyze the relationship the low friction floor was significantly smaller than
in data between RCOFmax and parameters with that on the high friction floor (Table 1). The
s i g n if i c an t d if f er e n c e s. F u rt h e r m o r e, f o r normalized step length on the low friction floor was
parameters showing a significant relationship, significantly smaller than that on the high friction
regression slopes were tested by linear regression floor. There was no significant difference in step
analysis using RCOFmax as an independent time between high and low friction floors. The
variable. P values less than 0.05 were considered normalized vertical displacement of the lateral
statistically significant. malleolus on the low friction floor was significantly
greater than that on the high friction floor. The
RESULTS length of time from heel contact to toe contact on
the low friction floor was significantly smaller than
The number of effective samples was 96 for both that on the high friction floor. On the low friction
high and low friction floors. The number of floor, the toe marker touched the floor before the
samples with at least one slip on the low friction heel marker in four of the twelve subjects. There
floor was 48 (50%) (slip group). Push off for the was no significant difference in normalized COPy
second step had the highest slip frequency (40%) max between high and low friction floors.
and the greatest slip displacement (101.0 ± 51.6 Since there were significant differences in
mm) (Fig. 3). The slip frequency and displacement normalized step length, normalized vertical
at heel strike for the first step were 16.7% and 27.4 displacement of the lateral malleolus and length of
± 34.6 mm, respectively. Slips did not occur during time from heel contact to toe contact between the
 37

Table 1. Comparison of kinetic and kinematic parameters at gait initiation

on high and low friction floors
parameter HFF LFF p-Value
RCOFmax 1.08 ± 0.12 0.63 ± 0.12 <0.0001
SL/Height 0.37 ± 0.04 0.31 ± 0.05 <0.001
ST (ms) 457.6 ± 54 454.9 ± 80.1 0.114
VDLM/Height 0.041 ± 0.015 0.049 ± 0.01 <0.001
H-TCT (ms) 48.4 ± 25.9 11.3 ± 55.7 <0.001
COPy max/foot size 0.023 ± 0.007 0.024 ± 0.03 0.289
Data are shown Mean ± S.D. RCOFmax: maximum required coefficient of
friction during push-off for the second step, SL/Height: normalized step
length of the first step, ST: step time for the first step, VDLH/Height: normal-
ized maximum vertical displacement of the lateral malleolus for the first step,
H-TCT: length of time between heel contact and toe contact for the first step,
COPy max/foot size: normalized maximum backward displacement for cen-
ter of pressure.

Table 2. Correlation between RCOFmax and Step Length/height or VDLM/

height or H-TCT on high and low friction floors
parameter HFF LFF
r p-Value r p-Value
SL/Height 0.586 <0.000 0.583 <0.0001
VDLM/Height 0.29 0.0284 0.107 0.3788
H-TCT (ms) 0.251 0.1041 0.052 0.7038
Correlation coefficient between normalized step length of the first step and maxi-
mum required coefficient of friction (Fx/Fz) (RCOFmax) during push-off for the
second step was 0.586 on the high friction floor and 0.583 on the low frection
floor, thus the correlation on both floors was significant (p<0.0001). Conven-
tions as Table 1.

high and low friction floors, the relationship were no significant differences in normalized step
between the RCOFmax of the support leg and each length, step time, normalized vertical displacement
of these three parameters was investigated. The of the lateral malleolus, length of time from heel
results showed that there was a significant contact to toe contact and normalized COPy-max
correlation between RCOFmax and normalized step between the slip and non-slip groups.
length on both high and low friction floors (Table
2). DISCUSSION
The results of a linear regression analysis showed
that, on both high and low friction floors, there was The results of this study indicate that, when
a s ign if ica nt po si tiv e co rre lat io n b et w ee n initiating gait on the low friction floor, the risk of
RCOFmax and the normalized step length (Fig. 4). slipping was greatest during push-off for the second
The correlations between RCOFmax and the two step. The reason for this is that the second step
other parameters, namely normalized vertical slipped during push-off when RCOF exceeded the
displacement of the lateral malleolus and length of static coefficient of friction. This study investigated
time from heel contact to toe contact, were not the relationship between RCOF and various
significant. movement parameters. The results showed that the
On the low friction floor, the RCOFmax for the movement of the first step was manipulated to
slip group (0.66 ± 0.07) was significantly greater reduce the step length in order to avoid increasing
than that for the non-slip group (0.60 ± 0.15). There RCOF.
38 J. Phys. Ther. Sci. Vol. 14, No. 1, 2002

movement is probably the quick generation of a

joint torque in the direction of plantar flexion at heel
strike. Therefore, it is suggested that the differences
of these two parameters, vertical displacement of
the lateral malleolus and the length of time from
heel contact to toe contact, between the low and
high friction floors are movement strategies in order
to avoid increasing RCOF at heel strike.
Many authors have pointed out that the gait of the
elderly can be characterized by reduced step length
and toe clearance (the maximum distance between
the floor and the toes in the swing phase), which
indicates that the movement of the legs is smaller
and reduced in the elderly22, 23). Nevertheless, the
results of the present study suggested that large leg
Fig. 4. Regression line for RCOFmax and step length/
movements are required when walking on a low
height on high and low friction floors. Step
Length/Height of HFF (filled circle) and LFF friction floor. Since RCOFmax was reduced when
(open circle) were plotted against RCOFmax. the young subjects in the present study walked on
The slope of the regression line on the high and the low friction floor, there is a possibility that the
low friction floors was 0.131 and 0.171, elderly are more likely to slip as their ability to
respectively, and thus as RCOFmax decreased, adjust RCOF is diminished11).
step length/height decreased significantly
(p<0.0001). A significant difference in RCOFmax was
observed between the slip and non-slip groups on
the low friction floor, thus indicating that
RCOFmax can be used as an essential parameter for
The present study showed a significant distinguishing between the slip and non-slip
correlation between RCOFmax during push off for conditions. However, no significant difference in
the second step and the step length for the first step step length was observed between the slip and non-
on the low friction floor (r=0.583). Bunterngchit et slip groups on the low friction floor. The reason for
al.19) suggested that there was a low but significant this is thought to be that the coefficient of static
correlation between RCOF at heel strike and step friction on the low friction floor was not constant.
length in steady gait (r=0.27). The reason for this Since the RCOFmax for the slip group ranged from
low correlation was that RCOF at heel strike is 0.547 to 0.776, the coefficient of static friction on
likely to be affected by foot velocity or heel the low friction floor should have been within this
deformation upon contact20, 21). To our knowledge, range. The main reasons for variations in the
the present study is the first to show a high coefficient of static friction could be attributed to
correlation between RCOF during push-off for the factors such as, an uneven thickness of the oil film,
second step and the step length of the first step. because some liquid paraffin remained attached to
However, as shown in Fig. 4, the RCOFmax on the the sole of the foot during each measurement, or
LFF is smaller than that of HFF at the same step sweat from the soles of the feet. Nevertheless, the
length. This suggests that factors other than the step variations in the coefficient of static friction did not
length, such as the ankle joint moment during push- affect the results of the present study.
off for the second step and the RCOF for the first
step, may be involved in the decrement of ACKNOWLEDGMENTS
RCOFmax.
Vertical displacement of the lateral malleolus on The authors wish to thank Professor K. Miyamoto
the low friction floor was significantly greater than for his thoughtful comments and suggestions
that on the high friction floor. In addition, the regarding the preparation of this manuscript. This
length of time from heel contact to toe contact on work was supported in part by Japanese Ministry, of
the low friction floor was smaller than that on the Education, Science and Culture (12832005).
high friction floor. The reason for this controlled
 39

REFERENCES responses to unexpected slips during walking in older

adults. J Gerontology. Medical Sciences, 1998, 53:
1) Nevitt MC: Falls in the elderly; risk factors and 471–480.
prevention. In: Gait disorders of aging; falls and 14) Bunterngchit Y, Lockhart T, Woldstad JC, et al.: Age
therapeutic strategies. New York: Lippincott-Raven related effects of transitional floor surfaces and
Press, 1997, pp. 13–36. obstruction of view on gait characteristics related to
2) Mano Y, Nakane R: The cause of gait disorders and slips and falls. International J Industrial Ergonomics,
falls in elderly persons. J Clin Reha, 1998, 7: 243–247 2000, 25: 223–232.
(In Japanese). 15) Brady RA, Pavol MJ, Owing TM, et al.: Foot
3) Norton R, Campbell J, Lee-Joe T, et al.: Circumstances displacement but not velocity predicts the outcome of
of falls resulting in hip fractures among older people. J a slip induced in young subjects while walking. J
American Geriatrics Society, 1997, 45: 1108–1112. Biomechanics, 2000, 33: 803–808.
4) Nyberg L, Gustafson Y, Berggren D, et al.: Falls 16) You JY, Chou YL, Lin CJ, et al.: Effects of slip on
leading to femoral neck fractures in lucid older people. movement of body center of mass relative to base of
J American Geriatrics Society, 1996, 44: 156–160. support. Clinical Biomechanics, 2001, 16: 167–173.
5) Overstall PW, Exton-Smith AN, Imms FJ, et al.: Falls 17) Ashley MJ, Gryfe CI, Amies A: A longitudinal study
in the elderly related to postural imbalance. British of falls in an elderly population II. Some circumstances
Medical Journal, 1977, 1: 261–264. of falling. Age Ageing, 1977, 6: 211–220.
6) Asaka T, Kurosawa K, Komuro H, et al.: A study on 18) Miller CA, Verstraete MC: A mechanical energy
the environmental factors preventing the physically analysis of gait initiation. Gait and Posture, 1999, 9:
disabled from going out. J Archit Plann Environ Eng, 158–166.
1995, 474: 83–90 (In Japanese). 19) Bunterngchit Y, Lockhart T, Woldstad JC, et al.: Age
7) Salkeld G, Cameron ID, Cumming RG, et al.: Quality related effects of transitional floor surfaces and
of life related to fear of falling and hip fracture in older obstruction of view on gait characteristics related to
women. BJM, 2000, 320: 341–346. slips and falls. International J Industrial Ergonomics,
8) Chaffin DB, Woldstad JC, Trujillo A: Floor/shoe slip 2000, 25: 223–232.
resistance measurement. American Industrial Hygiene 20) James DL: Rubbers and plastics in shoes and flooring.
Association, 1992, 53: 283–289. The importance of kinetic friction. Ergonomics, 1983,
9) Lanshammar H, Strandberg L: The dynamics of 26: 83–99.
slipping accidents. J Occup Accid, 1981, 3: 153–162. 21) Grönqvist R, Roine J, Korhonen E, et al.: Slip
10) Strandberg L: On accident analysis and slip-resistance resistance versus surface roughness of deck and other
measurement. Ergonomics, 1983, 26: 11–31. underfoot surfaces in slips. J Occup Accid, 1990, 13:
11) Lockhart TE: The ability of elderly people to traverse 291–302.
slippery walking surfaces. Proceedings of the Human 22) Hageman PA, Blanke DJ: Comparison of gait of young
Factors and Ergonomics Society 41st Annual Meeting, women and elderly women. Phys Ther, 1986, 66:
1997, 1: 125–129. 1382–1387.
12) Swensen E, Purswell J, Schlegel R, et al.: Coefficient 23) Winter DA, Patla AE, Frank JS, et al.: Biomechanical
of friction and subjective assessment of slippery work walking pattern changes in the fit and healthy elderly.
surfaces. Human Factors, 1992, 34: 67–77. Phys Ther, 1990, 70: 340–347.
13) Tang PF, Woollacott MH: Inefficient postural