Biotribology 5 (2016) 16–22

Contents lists available at ScienceDirect

Biotribology

journal homepage: http://www.elsevier.com/locate/biotri

Contribution of center of mass–center of pressure angle tangent to the

required coefficient of friction in the sagittal plane during straight walking
Takeshi Yamaguchi a,b,⁎, Kei Masani c,d
a
Graduate School of Engineering, Tohoku University, Sendai, Miyagi, Japan
b
Graduate School of Biomedical Engineering, Tohoku University, Sendai, Miyagi, Japan
c
Rehabilitation Engineering Laboratory, Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Ontario, Canada
d
Rehabilitation Engineering Laboratory, Lyndhurst Centre, Toronto Rehabilitation Institute—University Health Network, Toronto, Ontario, Canada

a r t i c l e i n f o a b s t r a c t

Article history: In human bipedal walking, the peak value of the traction coefficient (i.e., the ratio of the shear force component to
Received 9 September 2015 the vertical force component exerted on the floor) produced shortly after heel contact is termed as the required
Received in revised form 15 December 2015 coefficient of friction (RCOF). Based on a bipedal inverted pendulum model, with a whole-body center of mass
Accepted 22 December 2015
(COM) and reaction forces applied at the center of pressure (COP) of standing feet, RCOF in the sagittal plane
Available online 29 December 2015
(RCOFy) can be expressed as the sum of the tangent of the COM–COP angle and a residual term (RT), mainly com-
Keywords:
prising the moment around COM. In this study, we investigated the contribution of the tangent of the COM–COP
Friction angle to RCOFy during straight walking. The study involved four healthy young adult males. The participants were
Slip asked to walk on a 5-m long carpeted walkway. Each participant performed nine trials, i.e., three walking speeds
Gait (1, 1.4, and 1.9 m/s) × three step lengths (0.55, 0.75, and 0.95 m). COM was estimated using motion capture. COPs
Center of mass for the left and right feet were measured using eight force plates embedded in the walkway. RCOFy was calculated
Center of pressure from the anterior–posterior and vertical ground reaction force components measured using the force plates. We
found that the tangent of the COM–COP angle accounted for 91%–124% of the RCOFy value. This percentage
tended to decrease with increasing walking speed (p b 0.05). The magnitude of RT accounted for only 5.3%–
24% of RCOFy. These results suggest that the tangent of the COM–COP angle dominantly determines RCOFy.
© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction to sustain human locomotion [6–9]; thus, RCOF is an important factor

Slips are the most frequent events leading to falls at home and work-
place [1–3]. Slips occur if the tangential force applied to the floor during
walking reaches the friction force at the shoe–floor interface; slips do
not occur when the tangential force is less than the friction force during
the stance phase of the gait cycle. Thus, the ratio of the tangential force
to the vertical force applied to the floor, i.e., the traction coefficient,
must be less than the coefficient of friction at the shoe–floor interface
to prevent slipping [4]. The traction coefficient, calculated from the
ground reaction forces exerted between the shoe and floor while walk-
ing on a dry, uncontaminated surface, has been used to identify the lo-
cation during the gait cycle where slipping occurs. The required
coefficient of friction (RCOF) is the peak value of the traction coefficient
obtained at weight acceptance, which usually occurs shortly after heel
contact by the leading foot [5]. RCOF is recognized as the minimum co-
efficient of friction necessary at the interface between the shoe and floor
⁎ Corresponding author at: 6-6-01 Aramaki-Aza-Aoba, Aoba-ku, Sendai, Miyagi 980-
8579, Japan.
E-mail address: yamatake@gdl.mech.tohoku.ac.jp (T. Yamaguchi).
related to slipping during walking.
Therefore, research is increasingly focused on the relationship be-
tween RCOF and gait kinematics. Published studies [10–13] indicate
that RCOF is affected by walking speed, step length, and heel-contact ve-
locity. It was shown that the tangent of the angle between the leg and a
line perpendicular to the floor affects RCOF [6,14]. Burnfield and Powers
[15] asserted that the inclination angles of the line connecting the
whole-body center of mass (COM) to the center of pressure (COP;
termed the COM–COP angle) is a predictive kinematic variable of
RCOF during straight walking. Similarly, Yamaguchi et al. [8,12] demon-
strated that the tangent of the COM–COP angle strongly correlates with
RCOF during straight walking [12] as well as during turning, gait termi-
nation, and gait initiation [8]. Thus, the tangent of the COM–COP angle is
a good predictor of RCOF, the examination of which may improve our
understanding of the correlation between individuals' anthropometric
and gait characteristics and the onset of slipping. However, till date,
no study has expounded the theoretic rationale behind the high corre-
lation between the tangent of the COM–COP angle and RCOF. Further-
more, the quantitative contribution of the tangent of the COM–COP
angle to RCOF has not been investigated.

http://dx.doi.org/10.1016/j.biotri.2015.12.002
2352-5738/© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 T. Yamaguchi, K. Masani / Biotribology 5 (2016) 16–22 17

Fig. 2. Bipedal inverted pendulum model in the sagittal plane, with all mass at the center of
mass location and reaction forces applied at the center of pressure for leading and trailing
feet during a double-support period.

F z2 F y2 Mx
where RT ¼ tan θx2 þ − ; ð3Þ
F z1 F z1 F z1 zCOM

yCOP1 −yCOM
tan θx1 ¼ ; and ð4Þ
zCOM

yCOP2 −yCOM
tan θx2 ¼ ð5Þ
zCOM

Fig. 1. Typical time course of (a) vertical ground reaction forces (Fz); (b) anterior–
posterior ground reaction forces (Fy) for the leading foot (foot 1) and trailing foot
(foot 2); and (c) traction coefficient (Fy1/Fz1) for foot 1 from heel contact to toe off. DS
and SS represent double support and single support, respectively.
Fig. 1 shows the typical ground reaction force components and trac-
tion coefficient during the stance phase of the gait cycle (i.e., from heel
contact to toe off for a single foot) in the sagittal plane. As shown in
Fig. 1, the peak of the traction coefficient in the sagittal plane (RCOFy)
is produced in the final phase of the double-support period. To analyze
the effect of the tangent of the COM–COP angle on RCOFy, we assumed a
bipedal inverted pendulum model, with all mass at the COM location
and reaction forces applied at COP during double support (Fig. 2).
When we consider the sagittal plane with y and z coordinates, where
y is the horizontal axis and z is the vertical axis, the moment equation
is as follows:
Mx ¼ F z1 ðyCOP1 −yCOM Þ þ F z2 ðyCOP2 −yCOM Þ−F y1 zCOM þ F y2 zCOM
where Fy1 and Fy2 are the horizontal ground reaction forces of the lead-
ing and trailing feet, respectively; Fz1 and Fz2 are the vertical ground re-
action forces of the supporting feet; yCOM, yCOP1 and yCOP2 are the y
coordinates of COM and COPs of the supporting feet; Mx is the moment
around COM in the sagittal plane; and zCOM is the height of COM. The
traction coefficient for the leading foot in the sagittal plane (Fy1/Fz1) is
derived from Eq. (1) as follows:
F y1
¼ tan θx1 þ RT
F z1
where θx1 and θx2 are the angles of inclination of the lines connecting
COM to COPs of supporting feet.
Thus, in the sagittal plane, the traction coefficient for the leading foot
in the sagittal plane (Eq. (2)) is theoretically equivalent to the sum of
the tangent of the COM–COP angle of the leading foot (tanθx1) and the
residual term (RT). The strong correlation between the tangent of the
COM–COP angle and RCOF evident experimentally can be accounted
for by the assumption that the term of the tangent of the COM–COP
angle is dominant in Eq. (2).
Therefore, in this study, we aimed to test this assumption. Specifical-
ly, we quantitatively investigated the contribution of the tangent of the
COM–COP angle and RT in Eq (2). According to published studies
[10–13], the tangent of the COM–COP angle of the leading foot and RT
may be affected by walking conditions such as step length and gait
speed. Therefore, we controlled these factors and systematically investi-
gated the contribution of the tangent of the COM–COP angle to RCOFy.
2. Methods
2.1. Participants
ð1Þ
This study involved four healthy young adult males aged 24.3 ±
3.3 years (range: 22–29 years), 1.74 ± 5.3 m in height (range:
1.69–1.81 m), and weighing 69 ± 15.4 kg (range: 56–88 kg). Informed
consent was obtained from each participant.
2.2. Experimental procedures
Gait trials were performed on a 5-m long carpeted walkway. Eight
force plates (OR6-5; Advanced Mechanical Technology, Inc., Water-
town, MA, USA) were located approximately 2 m from the start position
ð2Þ
for collecting ground reaction forces, as shown in Fig. 3. An eight-
18 T. Yamaguchi, K. Masani / Biotribology 5 (2016) 16–22

Fig. 3. Schematic of the experimental set up of straight walking trials.

camera motion-capture system (Vicon 612; Oxford Metrics Group Plc., double support, respectively; Fz1 and Fz2 indicate the vertical ground re-
Oxford, UK) recorded three-dimensional (3D) motion data from 34 action forces for the leading foot and the trailing foot, respectively.
infrared-reflective markers attached bilaterally to the extremities and The y coordinate of COP for each foot was computed using the verti-
torso of the participants. Ground reaction forces and 3D motion data cal ground reaction components and COP positions calculated for each
were collected at a sampling rate of 120 Hz. The participants were pro- force plate, as follows:
vided with commercially available men's shoes with rubber outer soles
8
[Shore hardness score: 70 (A/15)]. >
> X4
yCOP i  f z i
>
> ; for the left foot
The participants were asked to walk in a straight line with their left < Fz
and right feet landing on left- and right-side force plates, respectively yCOP ¼ i¼1
ð8Þ
> Xy
8
>
> COP i  f z i
(see Fig. 3). The gait trial comprised nine blocks, i.e., three walking >
: ; for the right foot
F z
speeds (v = 1, 1.4, and 1.9 m/s) × three step lengths (L = 0.55, 0.75, i¼5

and 0.95 m). Step length was standardized by marks on the walkway,
and cadence was standardized using a metronome to regulate walking Note that, in this study, yCOP1 and yCOP2 indicate the y coordinate of
speed. Participants were asked to look straight ahead as possible as COP for the leading and trailing feet during double support, respectively.
they can. Each trial was replicated three times (i.e., 27 trials per partic- As we did not instruct to the subjects with which foot they should
ipant). The combination of test conditions, i.e., step length and walking start walk, the leading/trailing foot could be either of left or right foot.
speed, was randomized. The walking start-point was located 2 m before We determined leading or trailing foot by the time-series of vertical
the force plates, so that the ground reaction forces could be measured ground reaction force data (Fz) for the two consecutive steps. Thus, if
during steady-state rather than transient walking. the leading foot is the left foot, Fz1, Fy1 and yCOP1 are of left foot and
Fz2, Fy2 and yCOP2 are of right foot, respectively, and vice versa.
2.3. Data analysis The whole-body COM was computed from the kinematic data using
known Japanese segmental parameters [16] in 14-segment configura-
Kinetic and kinematic data were collected from each trial; thus, 24 tions comprising the head, torso, bilateral upper arms, forearms,
walks (two steps × three replications × four participants) were ana- hands, thighs, legs, and feet. The ground reaction force and 3D motion
lyzed in each trial. The anterior–posterior (fy_i) and vertical (fz_i) ground data were low-pass filtered (10 Hz) with a Butterworth dual-pass filter.
reaction forces for the left and right feet were collected separately by four The traction coefficient for the leading foot was calculated by divid-
left-side (i = 1–4) and four right-side (i = 5–8) force plates. COP for each ing Fy1 by Fz1. Although very large peak values for the traction coefficient
force plate was computed from the ground reaction force data using Vicon were recorded immediately after heel contact, these values are anoma-
software (Oxford Metrics Group Plc.). The resultant anterior–posterior lous because of division by small vertical forces and do not generate
and vertical ground reaction forces for the left- or right-foot Fy and Fz macroscopic slips contributing to falls [17]. To avoid selecting these
were calculated by combining the ground reaction forces obtained from values as RCOFy, the traction coefficient after the first 5% of the stance
the left- or right-side force plates, respectively, as follows: phase was analyzed [18].
The COM–COP angles in the sagittal plane (θx1 and θx2) are the incli-
8 4
>
> X nation of the line connecting COM to COPs for the leading and trailing
>
> f y i ; for the left foot
< feet in the sagittal plane, respectively, which were calculated using the
Fy ¼ i¼1
ð6Þ locations of COM, COPs, and vertical projection of COM on the floor ac-
>
> X
8
>
> f y i ; for the right foot cording to Eqs. (4) and (5). The value of tanθx1, RT and each term on
:
i¼5 the right side member of Eq. (3) at the instance of RCOFy [tanθx1_RCOFy,
RTRCOFy, Fy2/Fz1_RCOFy, (Fz2/Fz1)tanθx2_RCOFy, − Mx/(Fz1zCOM)_RCOFy, re-
8
>
> X
4 spectively] were used for the analysis.
>
> f z i ; for the left foot
<
Fz ¼ i¼1
ð7Þ 2.4. Statistical analysis
>
> X
8
>
> f z i ; for the right foot
:
i¼5 Statistical analysis was performed using SPSS Statistics for Windows
Version 19.0 (IBM Corp., Armonk, NY, USA). Bivariate regression analy-
Note that in this study, Fy1 and Fy2 indicate the anteroposterior sis between RCOFy and tanθx1_RCOFy was performed by grouping all trial
ground reaction forces for the leading foot and the trailing foot at a conditions to investigate the correlation between them. Two-way
 T. Yamaguchi, K. Masani / Biotribology 5 (2016) 16–22
repeated-measures analysis of variance (ANOVA) was used to test if the
mean values of RCOFy, tanθx1_RCOFy, and RTRCOFy were affected by step
length and walking speed. Two-way repeated-measures ANOVA was
also used to test if the percentage of tanθx1_RCOFy and RTRCOFy to RCOFy
values were affected by step length and walking speed. A post-hoc
paired t-test with a Bonferroni correction was used to determine specif-
ic significant differences between step length and walking speed. Signif-
icance levels were set at p = 0.05.
3. Results
Fig. 4 presents an example of the time courses from heel contact to
toe off of Fy1/Fz1, tanθx1, and RT. RCOFy appeared at 13% of the stance
phase, with a magnitude of 0.236. tanθx1 continuously decreased with
the percentage of the stance phase and became zero at 42% of the stance
phase, at which time the direction of anterior–posterior ground reaction
force changed from backward (braking force) to forward (propulsion
force). The tanθx1 at the instance of RCOFy (tanθx1_RCOFy) was 0.220.
On the other hand, the magnitude of RT was negatively large just after
the first heel contact and positively large just before toe off, whereas it
was very small at the instance of RCOFy (RTRCOFy = 0.016).
Fig. 5 shows the relationship between RCOFy and tanθx1_RCOFy ob-
tained from all trials. Bivariate regression analysis demonstrated that
RCOFy increased linearly with tanθx1_RCOFy and that RCOFy and
tanθx1_RCOF strongly correlate with each other (R = 0.935; R2 = 0.874;
p b 0.001).
Fig. 4. Representative time course of (a) traction coefficient (Fy1/Fz1); (b) center of mass–
center of pressure angle tangent (tanθx1); and (c) residual term (RT). Step length was
0.75 m and walking speed was 1.4 m/s.
19
Fig. 5. Relationship between the required coefficient of friction in the sagittal plane
(RCOFy) and the tangent of COM–COP angle (tanθx1_RCOFy) at RCOFy.
RCOFy, tanθx1_RCOFy, and RTRCOFy values for each step length and
walking speed are presented in Fig. 6. Two-way repeated-measures
ANOVA indicated that RCOFy values were not significantly affected by
walking speed or step length–walking speed interaction (p N 0.05);
however, RCOFy values were significantly affected by step length
(p b 0.001). A post-hoc analysis revealed that RCOFy values increased
with increasing step length for each walking speed (p b 0.01). In
contrast, two-way repeated-measures ANOVA demonstrated that
tanθx1_RCOFy values were significantly affected by step length (p b 0.001)
and walking speed (p b 0.05), but were not affected by step length–
walking speed interaction (p N 0.05). A post-hoc analysis indicated that
tanθx1_RCOFy values increased with increasing step length (p b 0.001)
and significantly decreased with increasing walking speed at a step length
of 0.55 m (p b 0.05). As shown in Fig. 6(c), the magnitude of RT for each
condition is significantly small; there was no clear systematic change in
RTRCOFy, and it was not significantly affected by step length or walking
speed (p N 0.05).
Fig. 7 shows the percentages of tanθx1_RCOFy and RTRCOFy to RCOFy for
each step length and walking speed. As shown in Fig. 7(a), tanθx1_RCOFy
accounted for 91%–124% of RCOFy. Two-way repeated-measures
ANOVA indicated that this percentage was not significantly affected by
step length and step length–walking speed interaction (p N 0.05);
however, it was significantly affected by walking speed (p b 0.05). A
post-hoc analysis revealed that this percentage was significantly lower
at a walking speed of 1.4 m/s than that at a walking speed of 1 m/s
(p b 0.01) when the step length was 0.75 m and that it was significantly
lower at a walking speed of 1.9 m/s than that at a walking speed of 1 m/s
when the step length was 0.95 m. As shown in Fig. 7(b), the contribu-
tion of RTRCOFy to RCOFy was significantly smaller than that of
tanθx1_RCOFy; the magnitude of the RTRCOFy value accounted for only
5.3%–24% of the RCOFy value. Two-way repeated-measures ANOVA in-
dicated that the contribution of RTRCOFy was not significantly affected
by step length and step length–walking speed interaction (p N 0.05);
however, it tended to be affected by walking speed (p = 0.05). A
post-hoc analysis revealed that its contribution at a walking speed of
1.4 m/s was significantly lower than that at 1 m/s (p b 0.01) when the
step length was 0.75 m, whereas its contribution at a walking speed of
1.9 m/s was significantly lower than that at 1 m/s when the step length
was 0.95 m.
RTRCOFy comprises three terms, as shown in Eq. (3). Fig. 8 shows
the relationship between the values of the first and second terms
in the right side member of Eq. (3) at the instance of RCOFy. This
clearly shows that these values have a strong negative correlation
(R = − 0.983, R2 = 0.967; p b 0.001) and plots a cluster around the
y = −x line, which means that these values almost cancel each other.
Thus, the third term in the right side member of Eq. (3), which indicates
the value of the moment around the whole-body COM divided by Fz and
20 T. Yamaguchi, K. Masani / Biotribology 5 (2016) 16–22

Fig. 6. Effect of step length and walking speed on the mean values of (a) the required coefficient of friction in the sagittal plane (RCOFy); (b) the tangent of COM–COP angle (tanθx1_RCOFy) at
RCOFy and (c) residual term (RTRCOFy) at RCOFy. Error bars represent standard deviation.

the height of COM, must dominantly contribute to RTRCOFy. As expected,
this moment-related term has a positive correlation with RCOFy, as
shown in Fig. 9 (R = 0.74, R2 = 0.55, p b 0.001).
4. Discussion
The bipedal inverted pendulum model suggests that the traction co-
efficient of the standing foot can be described by the sum of tanθx1 and
RT. Our results indicate that the magnitude of RCOFy during straight
walking is dominantly determined by the tangent of the COM–COP
angle in the sagittal plane. Furthermore, a shorter step length is effective
in reducing the tangent of the COM–COP angle at weight acceptance
during straight walking irrespective of walking speed, probably
resulting in reduced RCOFy values. Our results also suggest that walking
with a shorter step length and a higher walking speed reduces
the tangent of the COM–COP angle at weight acceptance, possibly
as a result of the more anterior position of COM compared with that
during slower walking. The contribution of the magnitude of the
moment-related term was significantly smaller than that of the tangent
of the COM–COP angle; however, it contributed a maximum of 24% to
RCOFy. The effect of the moment-related term on RCOFy was larger at
low walking speeds, implying that the difference (error) between

Fig. 7. Effect of step length and walking speed on the mean percentages of (a) the tangent of COM–COP angle (tanθx1_RCOFy) at RCOFy and (b) residual term (RTRCOFy) at RCOFy to RCOFy
values. Error bars represent standard deviation.
 T. Yamaguchi, K. Masani / Biotribology 5 (2016) 16–22
Fig. 8. Relationship between the first and second terms at RCOFy in Eq. (3).
RCOFy and the tangent of the COM–COP angle increases when walking
speed is low.
A fall is one of the main causes of life-threatening injuries in elderly
people [19,20]. Slip-induced falls account for a large percentage of the
total fall-related accidents among elderly people [21–23]. Elderly people
walk more slowly, with a shorter step length and a faster heel contact
velocity, compared with young adults [24,25], whereas the horizontal
COM velocity of elderly people is significantly slower than that of
young adults [26]. Despite these differences, the RCOF value of elderly
people is not significantly different from that of young adults [27,28].
Based on our results, this could be because of the combination of a
shorter step length and a slower horizontal COM velocity in elderly peo-
ple, which results in a tangent of the COM–COP angle during weight ac-
ceptance equivalent to that of young adults. Based on our model, a
weaker ankle-joint moment due to reduced ankle muscle strength in el-
derly people [28] could also have an impact on RCOFy because the
ankle-joint moment may affect the moment around the whole-body
COM at weight acceptance during straight walking (i.e., the contribution
of the moment-related term to RCOFy). Further studies are necessary to
investigate the effect of the tangent of the COM–COP angle and
moment-related term on the RCOF of elderly people.
In the present study, we only analyzed the contribution of the tan-
gent of COM–COP angle to the RCOFy value at weight acceptance. The
traction coefficient for the trailing foot can be expressed by the sum of
tanθx2 and the residual term. Previous studies indicated that the peak
value of traction coefficient at push-off phase correlates strongly with
the tangent of the COM–COP angle as well as at weight acceptance
phase [8,12]. Thus, it is assumed that the magnitude of the RCOFy at
push-off phase also be mainly determined by the tangent of the COM–
Fig. 9. Relationship between the values of the moment-related term and residual term
(RTRCOFy) at RCOFy.
21
COP angle similar to that at weight acceptance phase presented in this
study. Although a slip during push-off phase is not so hazardous during
straight walking as the COM is arrested within the base-of-support, we
definitely need to investigate the effect of the tangent of COM–COP
angle on the RCOFy at push-off phase in future work. This study involved
only young male participants. In future, it is necessary to investigate
whether our concept is applicable to gaits of both sexes at a range of
ages. In addition, we need to confirm if the tangent of the COM–COP
angle contributes to RCOF during other types of gait, such as turning,
gait initiation, and gait termination, typically used in daily life.
5. Conclusions
Our bipedal inverted pendulum model suggests that, in the sagittal
plane, the traction coefficient for the leading foot is expressed by the
sum of tanθx1 and RT. The results of this study indicate that the magni-
tude of RCOFy is mainly determined by the tangent of the COM–COP
angle in the sagittal plane during straight walking; the value of the tan-
gent of the COM–COP angle accounts for 91%–124% of the RCOFy value.
Conversely, the contribution of the RT, mainly comprising the moment-
related term, was significantly smaller than that of the tangent of the
COM–COP angle, although its contribution slightly increased with a de-
crease in walking speed. These results suggest that walking with a
smaller COM–COP angle tangent at weight acceptance effectively re-
duces the RCOFy value, i.e., the risk of slipping, during straight walking.
Walking with a shorter step length significantly decreased the tangent
of the COM–COP angle at weight acceptance, resulting in a lower
RCOFy value. Further investigations are necessary to identify whether
our findings are applicable to other types of gait and other populations,
such as women and individuals of different ages.
Conflict of interest statement
No author of this paper has any conflict of interest, including specific
financial interests, relationships, and/or affiliations relevant to the sub-
ject matter or materials included in this manuscript.
Acknowledgments
This study was partially supported by Grant-in-Aid for Scientific
Research (C) (25420080).
References
[1] T.K. Courtney, G.S. Sorock, D.P. Manning, J.W. Collins, M.A. Holbein-Jenny,
Occupational slip, trip, and fall-related injuries -can the contribution of slipperiness
be isolated? Ergonomics 44 (2001) 1118–1137.
[2] M.H. Choi, H.S. Kim, B. Kim, J.C. Lee, S.J. Park, U.H. Jeong, et al., Extraction and anal-
ysis of risk elements for Korean homecare patients with senile dementia, J Behav
Health Serv Res 43 (2016) 116–126.
[3] G.A. Koepp, B.J. Snedden, J.A. Levine, Workplace slip, trip and fall injuries and obe-
sity, Ergonomics 58 (2014) 674–679.
[4] T. Yamaguchi, K. Hokkirigawa, “Walking-mode maps” based on slip/non-slip
criteria, Ind Health 46 (2008) 23–31.
[5] M.S. Redfern, R.O. Andres, The analysis of dynamic pushing and pulling; required co-
efficients of friction, Proceedings of the 1984 International Conference on Occupa-
tional Ergonomics, I 1984, pp. 573–577.
[6] P.J. Perkins, Measurement of slip between the shoe and ground during walking, in:
C. Anderson, J. Senne (Eds.), ASTM STP 649. Walkway Surfaces: Measurement of Slip
Resistance, ASTM, Philadelphia 1978, pp. 71–87.
[7] L. Strandberg, H. Lanshammar, The dynamics of slipping accidents, J Occup Accid 3
(1981) 153–162.
[8] T. Yamaguchi, M. Yano, H. Onodera, K. Hokkirigawa, Kinematics of center of mass
and center of pressure predict friction requirement at shoe–floor interface during
walking, Gait Posture 38 (2013) 209–214.
[9] P. Fino, T.E. Lockhart, Required coefficient of friction during turning at self-selected
slow, normal, and fast walking speeds, J Biomech 47 (2014) 1395–1400.
[10] S. Carsöö, A method for studying walking on different surfaces, Ergonomics 5 (1962)
271–274.
[11] R.W. Soames, R.P.S. Richardson, Stride length and cadence: their influence on
ground reaction forces during gait, in: D.A. Winter, et al., (Eds.), Biomechanics IX,
Human Kinetics Publisher, Champaign, IL 1985, pp. 406–410.
22 T. Yamaguchi, K. Masani / Biotribology 5 (2016) 16–22
[12] T. Yamaguchi, S. Hatanaka, K. Hokkirigawa, Effect of step length and walking speed
on traction coefficient and slip between shoe sole and walkway, Tribol Online 3
(2008) 59–64.
[13] T.E. Lockhart, J.C. Woldstad, J.L. Smith, Effects of age-related gait changes on the bio-
mechanics of slips and falls, Ergonomics 46 (2003) 1136–1160.
[14] R. Grönqvist, J. Roine, E. Jarvinen, E. Korhonen, An apparatus and a method for deter-
mining the slip resistance of shoes and floors by simulation of human foot motions,
Ergonomics 32 (1989) 979–995.
[15] J.M. Burnfield, C.M. Powers, The role of center of mass kinematics in predicting peak
utilized coefficient of friction during walking, J Forensic Sci 52 (2007) 1328–1333.
[16] M. Ae, H.-P. Tang, T. Yokoi, Estimation of inertia properties of the body segment in
Japanese athlete, J SOBIM 11 (1992) 23–33 (in Japanese).
[17] M.S. Redfern, R. Cham, K. Gielo-Perczak, R. Grönqvist, M. Hirvonen, H. Lanshammar,
et al., Biomechanics of slips, Ergonomics 44 (2001) 1138–1166.
[18] M.G. Blanchette, J.R. Brault, C.M. Powers, The influence of heel height on utilized co-
efficient of friction during walking, Gait Posture 34 (2011) 107–110.
[19] M.W. Rogers, M.L. Mille, Lateral stability and falls in older people, Exerc Sport Sci Rev
31 (2003) 182–187.
[20] L. Calandre, I. Conde, P.F. Bermejo, Gait and stability disorders of the elderly. Clinical
analysis of a series of 259 patients older than 70 years, Neurologia 20 (2005)
232–239.
[21] A.J. Campbell, J. Reinken, B.C. Allan, G.S. Martinez, Falls in old age: a study of fre-
quency and related factors, Age Ageing 10 (1981) 264–270.
[22] L.Z. Rubenstein, A.S. Robbins, B.L. Schulman, J. Rosado, D. Osterweil, K.R. Josephson,
Falls and instability in the elderly, J Am Geriatr Soc 36 (1988) 266–278.
[23] J. Agnew, A.J. Suruda, Age and fatal work-related falls, Hum Factors 35 (1993)
731–736.
[24] D.A. Winter, A.E. Patla, J.S. Frank, S.E. Walt, Biomechanical walking pattern changes
in the fit and healthy elderly, Phys Ther 70 (1990) 340–347.
[25] T.E. Lockhart, The ability of elderly people to traverse slippery walking surfaces, Pro-
ceedings of the Human Factors and Ergonomics Society 41st Annual Meeting,
Human Factors and Ergonomics Society, Santa Monica, CA 1997, pp. 125–129.
[26] T.E. Lockhart, J.C. Woldstad, J.L. Smith, Effects of age-related gait changes on the bio-
mechanics of slips and falls, Ergonomics 46 (2003) 1136–1160.
[27] T.E. Lockhart, J.L. Smith, J.C. Woldstad, Effects of aging on the biomechanics of slips
and falls, Hum Factors 47 (2005) 708–729.
[28] L. Wolfson, R. Whipple, P. Amerman, J. Kaplan, A. Kleinberg, Gait and balance in the
elderly, Clin Geriatr Med 1 (1985) 649–659.