Biomechanical analysis of sit-to-stand transfer in healthy and

paraplegic subjects
Fariba Bahrami
a,
*
, Robert Riener
b
, Parviz Jabedar-Maralani
a
, G unther Schmidt
b
a
Department of Electrical and Computer Engineering, Faculty of Engineering, Building No. 2, North Kargar Avenue, Tehran University, Tehran 14399,
Iran
b
Institute of Automatic Control Engineering, Technical University of Munich, Munich, Germany
Received 12 March 1998; accepted 18 May 1999
Abstract
Objective. An experimental study of the sit-to-stand transfer in healthy adults with/without arm-support and in paraplegic pa-
tients with/without electrical stimulation of the quadriceps muscles was performed. The study was aimed to compare the joint
torques, momentum transfer hypothesis, and stability of the sit-to-stand transfer in the healthy and paraplegic subjects.
Methods. A planar 3-linkage rigid body model was used to compute the body-segmental linear momentum and the reaction
forces and torques at the joints from measured data.
Results. In healthy subjects the arm-support enlarged the support base of the body and thus, increased the postural stability.
Strong arm-assistance reduced the maximum hip and knee joint torques by more than 50%. It was observed that the healthy
participants rising with arm-support used momentum transfer to facilitate the transition from sitting to standing. The paraplegic
participants did not apply the momentum transfer strategy and the sit-to-stand transfer was accomplished in a quasi-static manner.
Stimulating the quadriceps, the legs could participate partly in the movement dynamics.
Conclusion. Our results indicate that some signicant dierences exist between the maneuver applied by the paraplegic patients to
stand up and the strategies used by the healthy adults rising with arm-support.
Relevance
Analysis of the biomechanical factors underlying the sit-to-stand activity is essential in the design of competent closed-loop
neuroprosthesis controllers which assist paraplegic patients during rising. 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Biomechanics; Standing up; Spinal cord injury; Functional electrical stimulation
1. Introduction
Sit-to-Stand (STS) transfer is one of the most com-
mon daily movements, which is also a pre-requisite for
many other activities. STS transfer is mechanically a
demanding task [1]. It requires adequate torques be
developed at each joint, while spatial and temporal
motion of the body segments are coordinated. It has
been demonstrated that persons with spinal cord injuries
can re-obtain the capability of standing up [2,3], stan-
ding [35] and walking [6] by means of Functional
Electrical Stimulation (FES) of the lower extremity
muscles. Although FES is a useful and feasible method,
there are still many problems which must be solved to
enhance the eciency of the FES-controllers [3,7]. For
example, in most FES systems the voluntary movement
of the upper body (intact limbs) causes external forces
and torques disturbing the FES-induced motion of the
paralyzed lower extremities. To enhance the perfor-
mance of the FES-controllers, it has been proposed to
integrate the eect of the voluntary movement of the
trunk and arm-support in the controller [810]. Biome-
chanical analysis of the factors which describe the ability
or disability of a given movement (e.g., standing up) in
dierent healthy and paraplegic subjects, will lead us to
nd an eective solution for the last posed problem.
Many previous studies have analysed dierent fun-
damental aspects and biomechanical factors of the STS
movement. Coghlin and MacFadyen [11] have reported
that two dierent strategies were used by their test
Clinical Biomechanics 15 (2000) 123133
www.elsevier.com/locate/clinbiomech
*
Corresponding author.
E-mail address: bahramy@sofe.ece.ut.ac.ir (F. Bahrami).
0268-0033/00/$ - see front matter 2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 6 8 - 0 0 3 3 ( 9 9 ) 0 0 0 4 4 - 3
groups to stand up without arm-support. They called
these schemes knee and hip-trunk strategies. In the hip-
trunk strategy the hip and ankle joint torques are greater
than the corresponding joint torques in the knee strat-
egy. Thus, Doorenbosch et al. [12] have suggested that,
the persons with muscle weakness of the knee extensors
should use additional compensatory mechanisms, like
the arm-assistance. Applying a semi-dynamic calcula-
tion, Arborelius et al. [13] have shown that the arm-
support during standing up reduces the mean maximum
hip joint torque by 50%. Schultz et al. [14] have com-
pared the joint torque requirements at the seat-o in-
stant with the maximum voluntary joint strengths.
Whence, they have inferred that except for the people
who have some disabilities in their lower extremities,
joint torque requirements may not be the only factor
limiting the ability to lift o. Riley et al. [1] have eval-
uated two possible strategies in healthy adults during
standing up without arm-support. Their results indicate
that, the upper body exion before leaving the seat is
either to garantee the postural balance just after seat-o
(strategy I) or it is to generate a suitable horizontal
linear momentum (HLM). The HLM is then trans-
formed to the vertical linear momentum (VLM) to fa-
cilitate the dynamic transition from sitting to standing
(strategy II). In another study Riley et al. [15] have
shown that the sit-back failures occur by insucient
momentum and torque generation, and justied the role
of momentum control in a stable STS transfer without
arm-support.
However, to the best of our knowledge, few com-
prehensive analysis of the arm-supported STS transfer
in healthy adults has been accomplished. In addition,
there are few works on the voluntary functions of the
trunk and arms during STS transfer in paraplegic pa-
tients [10,16]. Given the existing gaps, this study has
been conducted with the following purposes in mind:
(a) Accomplishing a dynamic analysis of arm-support-
ed chair rise in healthy adults in order:
(a.1) to evaluate the stability of arm-supported
standing up;
(a.2) to investigate the momentum control hypoth-
esis; and
(a.3) to study the eect of arm-support in the joint
torque distributions during STS transfer.
(b) Investigating the function of the intact limbs in
the paraplegic patients during STS transfer without/
with FES and to nd out if the paraplegic subjects ap-
ply also the momentum control strategy to stand up.
According to Riley et al. [1,15,17] the displacements
of the Center of Pressure (CoP) under the support base
of the feet and the body Center of Mass (CoM) relative
to the ankle may be used as criteria for dynamic (pos-
tural stability) and static stability (postural balance) of
the movement, respectively. Thus, the time history of the
spatial displacement of the body CoM, the CoP under
the feet, and the LMs will be used to analyze the stability
of the movement. The horizontal and vertical LMs and
the kinetic energy will clarify the momentum transfer
hypothesis. Further, the vertical ground reaction forces,
the joint torques, and the forces at the shoulder will
explain how the healthy and paraplegic participants
distributed the movement loads between their hands
(shoulders) and legs during STS transfer.
2. Methods
2.1. Subjects
The experimental data were collected from 10 healthy
participants and two paraplegic patients with lesions at
T8 level. Informed consent was obtained for each sub-
ject prior to data collection. The age, sex, body weight
and height, and segmental lengths of all subjects are
listed in Table 1. The mean and standard deviation
values of the anthropometric data for healthy partici-
pants are also given in the same table. Shank length was
measured from the medial malleolus to the visually de-
termined approximate knee center of rotation; thigh
length was measured from the knee center of rotation to
the great trochanter, and the length of the trunk was
measured from femur to the center of rotation of the
shoulder, to which the lengths of neck and head were
later added to obtain the length of the upper body.
Masses, mass centers and inertia tensors for each seg-
ment were calculated from regression equations [18],
and scaled for each subject to her/his body weight (BW)
and body height (BH).
2.2. Protocols
All subjects were asked to sit comfortably in a self-
selected body state on a chair without back and arm-
rests, while keeping their back upright. Both feet were
placed symmetrically and parallel to each other on a
forceplate. To stand up with arm-support, all partici-
pants have used two xed bars located in front of them.
Locating the bars in front of the healthy participants,
enlarged considerably the support base of the body. In
our measurements Mean (X
wrist
X
ankle
) was 27.97 cm
(SD, 4.5 cm). For the paraplegic participants to be able
to rise, the seat had to be located almost between two
bars such that Mean (X
wrist
X
ankle
) =4.8 cm (SD, 4.1
cm). The height of the chair and bars as well as the
distance between the chair and bars have been adjusted
for each subject separately so that the subjects could rise
from the chair in a natural and comfortable manner
while the feet were completely kept on the force plate
and the hands were holding the bars (Fig. 1). It was
encountered that the average preferred height of the
chair and bars were 99.4% of subject's shank length (SD,
124 F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133
7.6% SL) and 50.58% of the subjects' BH (SD, 2.02%
BH), respectively.
The healthy subjects were asked to stand up in 3 dif-
ferent manners and then remain motionless: (a) 2 times
with crossed arms on the chest, (b) 5 times with normal
assistance of the arms, and (c) 5 times with strong arm-
support. In all cases the subjects were instructed to rise at
their own natural speed and the way they felt would be
the most usual manner for that situation. In case (a) the
arms should not move. In case (b) the function of arms
was principally to support the stability of the movement.
In case (c) the relative distance between chair and bars
for each subject was the same as in case (b). In case (c) the
subjects were asked to use their hands as the main sup-
port of the body and to regard their legs to be too fa-
tigued. For each trial 4 s of data were collected. The
paraplegic patients have performed the STS task: (a) 5
times only with the help of their arms, and (b) 5 times
with additional surface stimulation of quadriceps mus-
cles. A two-channel stimulator was used and the stimu-
lating pulses were modulated by a ramp signal with a
duty-cycle of 2 s. For each paraplegic participant 9 s of
data have been recorded.
2.3. Data and equipment
Infrared-reecting markers were xed on the ankle,
knee, hip, shoulder, elbow, and wrist joints from the
right side of the body, and the 3-dimensional (3D)
Cartesian coordinates of each joint (with an accuracy of
1 mm for position) have been recorded by two CCD
cameras of an ELITE system (BTS, Milan, Italy). The
vertical seat reaction force (SRF) was obtained from an
instrumented laboratory chair. A Kistler piezoelectric
force plate type 9284 (Winterthur, Switzerland) was
used to measure the 3D ground reaction forces (GRF)
under both feet (with an accuracy of 1% of the full
Table 1
Data about the ten healthy adults and two paraplegic subjects participated in the measurements. (In the table, HA stands for Healthy Adults)
No. Name Sex Diag-
nosis
Age Weight
[kg]
Height
[cm]
Trunk
[cm]
Fore-
arm
[cm]
Upper
arm
[cm]
Thigh
[cm]
Shank
[cm]
Foot
[cm]
Height
of
chair
[cm]
Height
of bars
[cm]
1 MM f T8 39 46 158 56 25 30 40 40 24 36 74
2 UK m T8 30 74 183 58 32 30 46 50 30 45 88
3 BR f HA 26 60 169 51 25.5 31 44.5 45.5 24.5 40.5 84
4 HH m HA 33 74 187 55 30 34 46.5 50 28 49.5 95.5
5 II f HA 27 64 164 53.5 26.5 27.5 39.5 43.5 24 42.5 86
6 JD m HA 26 65 181 55 30.5 33.5 45 49.5 26 49.5 95.5
7 OR m HA 31 81 180 53.5 27 29.5 42 42 27 49.5 94.5
8 RR m HA 28 83 185 53.5 29 31.5 44 47.5 26 49.5 94.5
9 SE f HA 26 65 165 52 26 31.5 41.5 43.5 24 40.5 85.5
10 SV m HA 27 72 184 55 30 33 43 50 29 49.5 94.5
11 TE m HA 29 75 180 56.5 30 34.5 41.5 44.5 26 42.5 88.5
12 UW m HA 26 70 175 55 29 32.5 39 47 26 46.5 93.5
Mean
(HA)
29 70.9 177 53.98 28.35 31.85 42.65 46.3 26.05 46 91.2
SD 3.88 7.55 8.35 1.64 1.9 2.15 2.4 2.93 1.64 4.03 4.64
Fig. 1. Schematic representation of the experimental set-up used
during the measurements. Since the movement was assumed to be
symmetrical, a 3DoF planar rigid body model has been used to des-
cribe the dynamics of the STS transfer in the healthy and paraplegic
subjects. The origin of the coordinate system was assumed to be on the
ankle joint. The positive directions for joint angles, displacement and
velocity of dierent components of the CoM and LM as well as re-
action forces are according to the coordinate system shown in the
gure.
F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133 125
scale) and the corresponding CoP. The 3D reaction
forces under each hand (HRF) were measured with two
laboratory bars instrumented with strain gauges. All
kinematic and kinetic data have been recorded simul-
taneously and sampled at 100 Hz. The measured data
were ltered with a 9th order forwardreverse Butter-
worth lter with 4 Hz cut-o frequency [19].
2.4. Model
The movement was assumed to be symmetrical [20].
The head and neck were also assumed to remain along
the trunk. Thus, a 3-linked rigid body model (see Fig. 1)
with 3 degrees of freedom (DoF) was used to describe
the dynamics of the upper body and lower extremities in
the sagittal plane [14,21]. In arm-supported standing up
the function of arms has been integrated into the model
as external loads at the shoulders which could vary with
time during the movement. It has been assumed that the
chair exerts forces at the hip joint onto the femur. The
chair reaction forces were described by its horizontal
and vertical components. Thus, two nonlinear spring
dashpot pairs were used to model the visco-elastic
characteristics of bodychair contact in each direction.
The parameters of the springdashpot pairs have been
identied for each subject separately so that the forces
applied by the chair become zero at the seat-o instant.
Some of the previous studies have used only quasi-
dynamic models to evaluate the total joint forces and
torques [13,14]. According to the more recent analysis of
Hutchinson et al. [22] although the static loads dominate
the joint forces and torques, the dynamic analysis can
especially improve the accuracy of the calculations at the
upper-most joints like back and shoulders. Thus, the
dynamic equations of the system have been recursively
derived by applying NewtonEuler equations [23,24] to
each link starting from the feet:
s
ankle
= m
foot
g l
mfoot
CoP p
ygr
Y
w(q) q g(qY q) q(q)
= t
/
w
s
T
w
sh
2 3
t
/
ch
p
chair
t
/
gr
p
ground
Y
(1)
p
xsh
= m
shank

xshank
m
thigh

xthigh
m
trunk

xtrunk
p
xgr
p
xch
Y
p
ysh
= m
shank
(
yshank
g) m
thigh
(
ythigh
g)
m
trunk
(
ytrunk
g) p
ygr
p
ych
Y
(2)
where qY q and q are relative joint angular displacement,
velocity and acceleration vectors, M(q) the inertia ma-
trix of the system, g(qY q) the term comprising the co-
riolis and centrifugal forces, G(q) the gravitational term,
s
T
the total joint torque vector at ankle, knee and hip
joints, M
sh
, F
xsh
and F
y sh
the reaction torques and forces
(horizontal and vertical) at the shoulder, respectively,
and J
T
, J
ch
and J
gr
Jacobians, mapping the external
torques and forces to the joint space; m, l
m
, a
x
and a
y
are
segmental masses, distances of segmental CoM from
distal joints and horizontal and vertical velocities of
segmental CoMs, respectively. F
chair
is the vector of ex-
ternal forces exerted by the chair at the hip joint in the
horizontal (X) and vertical (Y) directions according to
the following mathematical description:
p
ich
= u
ich
h
f
ich

l
i
i
e
[li [

Y i = Y (3)
where l
i
is the incremental change of the spring length in
X or Y direction; K
ich
and B
ich
are the stiness and
damping parameters of the springdashpot pairs, re-
spectively.
2.5. Treatment of data
Applying the sagittal plane components of the mea-
sured kinematic and kinetic data to the 2D/3DoF rigid
body model, the relative joint angles, body CoM, hori-
zontal and vertical LM of the body CoM together with
the joint torques, forces and the external loads at the
shoulder have been computed. To calculate the angular
velocity and acceleration for each joint, a fth order
polynomial has been tted to each ve consecutive
samples of the joint angle.
To allow comparisons, the collected and calculated
data have been normalized according to the following
procedure. The total duration of the movement has been
dened to be 100, which began at the instant of forward
movement of the trunk until the moment when standing
fully upright. The CoP and X
CoM
have been normalized
to the units of percent of the foot length, and the vertical
displacement of the CoM (Y
CoM
) to the body height
(%BH). The horizontal and vertical LMs together with
the reaction forces have been normalized to the units of
percent of the body weight (%BW), and torques to the
units of percent of the BW BH. The same normaliza-
tion techniques for the kinematic and kinetic data are
also suggested by Hof [25]. Later the results have been
averaged for all the healthy subjects. For each paraple-
gic patient participating in this study, only two typical
trials from ve performed tasks in each case have been
chosen, and then the results have been averaged and
normalized. The criteria for trial selection were: (1) none
of the markers were hidden from any of the CCD
cameras at any moment, (2) the feet remained xed on
the force plate all the time during rising, and (3) no
spasticity occurred during the motion.
3. Results
Table 2 lists the total duration of the STS tasks per-
formed by the healthy and paraplegic participants with
126 F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133
dierent strategies (Mean values and standard devia-
tions). The total duration of the movement is divided
into two phases. Phase I is before leaving the chair, and
phase II begins after seat-o until the body reaches the
standing posture. Figs. 26 depict the results. The pos-
itive torques act to extend the joints and the numerical
values are calculated for one joint. In Fig. 6 the knee
joint angle is plotted according to the AAOS system
[26]. The full extension of the knee joint is dened as 0

and the negative values indicate the exion of knee. The

time at which the healthy subjects left the seat was de-
termined as the instant when the horizontal component
of the GRF attained its minimum value [21] and named
seat-o instant. It was not possible to apply the same
denition to determine the exact seat-o event for the
paraplegic participants. For the paraplegic subjects the
seat-o was investigated directly as the instant at which
the body left the contact with the seat. The joint coor-
dinates in the Cartesian space as well as the joint angles
and seat reaction forces were used to determine the
approximate seat-o instant for the two paraplegic
participants. In rising with or without FES, the seat-o
event occurred almost at the same relative time
(when almost 42% of the movement duration is over). In
the gures the seat-o instants for the healthy subjects
are marked by vertical lines of the line style corre-
sponding to the main curves; the approximate seat-o
instants for the paraplegic patients are also shown by
vertical lines.
3.1. Healthy adults
3.1.1. Body center of mass
According to Fig. 2, in rising with arm-support the
horizontal and vertical displacement of CoM occurred
simultaneously from the beginning of the movement. In
rising with normal arm-support leaving the seat, the
body CoM was anterior to the ankle, outside of the
support base of the feet (10.75 cm, SD=4 cm). In
standing up with strong arm-support, the X
CoM
at seat-
o was located anterior to the ankle (4.1 cm, SD=8
cm), under the heel (Mean (X
heel
X
ankle
) =5.7 cm,
SD=0.67 cm). The relative high s.d.s for X
CoM
X
ankle
in the STS transfer with strong arm-support may be due
to the subjective preferences in establishing postural
stability when leaving the seat. It was observed that in
all cases the CoP at seat-o instant was under the sup-
port bases of the feet and the CoM was displaced from
Fig. 2. Normalized and averaged spatial displacement (XY projec-
tion) of the body Center of Mass (CoM) for the healthy adults (HA)
and paraplegic patients (PP). In each case the horizontal projection of
the CoM at the seat-o instant is also marked.
Table 2
Average duration (standard deviation) of the STS transfer for dierent modes of the task accomplishment
a
(a) Healthy adults
Without arm-support With normal arm-support With strong arm-support
Total Phase I Phase II Total Phase I Phase II Total Phase I Phase II
Duration [s] 1.738 0.706 1.032 1.956 0.707 1.249 2.801 1.085 1.716
(0.46) (0.07) (0.07) (0.34) (0.10) (0.10) (0.67) (0.35) (0.35)
Duration [%] 100 40.63 59.37 100 36.125 63.875 100 38.75 61.25
(b) Paraplegic patients
Without FES With FES
Duration [s] 4.25 (0.57) 5.175 (1.4)
a
For healthy adults phases I and II are represented as absolute time and as percentage of the total duration of the movement. For the paraplegic
participants it was not possible to determine the exact seat-o instant.
F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133 127
the CoP. Table 3 summarizes the average horizontal
distance of the ankle and of the body CoM at seat-o
instant from CoP.
3.1.2. Linear momentum
In standing up with arm-support, the horizontal and
vertical LM were almost everywhere positive and the
maximum HLM occurred about the seat-o event (Fig.
3). The HLM at seat-o was always positive, and in
rising without arm-support it was larger
(HLM[
seat-off
=30%BW cm/s) than in two other strate-
gies (HLM[
seat-off
20%BW cm/s). On the other hands,
leaving the chair with normal arm-support the VLM
was larger than (20%BW cm/s) VLM in two other cases
(less than 10%BW cm/s). Riley et al. [1] observed that
the kinetic energy of the CoM had two peaks which
occurred almost when HLM and VLM achieved their
maximum values. They dened the time interval be-
tween maximum HLM and maximum VLM as Mo-
mentum Transfer Phase (MTP). According to our results
(Fig. 4), in rising with arm-support, the kinetic energy of
the CoM had also two local maxima which occurred
almost at the same instants as the HLM and VLM
Fig. 3. Normalized horizontal and vertical linear momentum of the
body CoM for the healthy adults (HA) and paraplegic patients (PP).
Fig. 4. Averaged kinetic energy of the body CoM for healthy adults
and paraplegic patients. For each case the two successive peaks of the
kinetic energy are marked with vertical lines.
128 F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133
attained their maximum values, but these peaks were
not as large and distinct as in the STS transfer without
arm-support. The vertical and horizontal LMs in STS
transfer without arm-support have larger maximum
values than those maximum values in the STS with arm-
support. In addition, the MTPs for rising with normal
and strong arm-support (0.14 and 0.28 s, respectively)
were shorter than MTP in standing up without arm
support (0.522 s).
3.1.3. Reaction forces
According to our results, for standing up without
arm-support the vertical GRF (VGRF) had a larger
peak (113%BW, SD=10.8%BW) compared with the
VGRF in the other strategies. In rising with normal
arm-support, the VGRF had a smaller peak (104%BW,
SD=6.8%BW) than the VGRF in the STS without arm-
support. Since in the former, short after seat-o the
velocity of the body ascension was began to decrease. In
standing up with strong arm-support, the bars carried a
part of the subject's BW. However, the positive peak in
the VGRF (68%BW, SD=11%BW) occurred shortly
after seat-o, shows that the legs shared the dynamics of
the movement with hands.
3.1.4. Joint torques
According to Table 4, in rising with normal/strong
arm-support maximum torques at the knee and hip
joints were smaller (up to 50%) than in the STS transfer
without arm-support, but applying normal arm-support
the ankle joint torque was relatively larger than the
ankle joint torque during STS movement without arm-
assistance. In all cases the total knee and hip joint tor-
ques attained their maximum shortly before seat-o.
3.1.5. Shoulder loads
Table 4 summarizes also the average maximum values
of the torques and forces at one shoulder for healthy
Fig. 5. Normalized average external loads at shoulder for healthy
adults and paraplegic patients.
Fig. 6. Knee joint angles averaged over the data of the two paraplegic
patients.
F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133 129
subjects rising with arm-support. Fig. 5 shows that in
the STS transfer with arm-support the vertical forces at
shoulder assisted the upward movement of the body. It
was mentioned that in rising with normal arm-support,
the subjects lifted from the chair before locating the
horizontal projection of the CoM under the feet (Fig. 2).
The horizontal forces at shoulder together with larger
torque at ankle supported the body in this case and
prevented it from sit-back failure. In standing up with
strong arm-support, the negative external torques at the
shoulder were mainly due to the rotation of the trunk
about the shoulder joint. The positive horizontal force
(Fig. 5) at the shoulder indicates that applying strong
arm-support, the subjects pulled the bars with a con-
siderable horizontal force (maximum 18%BW) to move
forward and to locate their trunk between two hands.
This maneuver facilitated separation from the seat and
reduced notably the maximum forces and torques at the
lower extremity joints about seat-o instant (up to 50%).
3.2. Paraplegic patients
In standing up without FES the patients launched
their body forward (see Fig. 6) to be able to leave the
chair (launching maneuver). In the FES-assisted STS
transfer the patients did not use the launching maneuver
to leave the seat. Moreover, at the end of phase II the
knee joint was more extended than in the standing up
without FES.
3.2.1. Body center of mass
The spatial displacement of the CoM (Fig. 2) indi-
cates that before leaving the chair, the paraplegic pa-
tients located their body CoM near or between two
hands. Launching maneuver implied the additional
downward displacement of the CoM in the STS transfer
without FES. After establishing the CoM in the support
base of the body, the patients rose almost vertically
trying to keep the trunk between two hands.
3.2.2. Linear momentum
Fig. 3 shows that, in standing up without/with FES
the VLM about seat-o was almost zero. In rising
without FES the positive peaks in the HLM (20%BW
cm/s) about and short after seat-o are mainly due to
the launching maneuver. After standing up, the patients
could not stand properly and the arms and trunk ut-
tered, whence, the HLM at the end of movement was
not zero. In the FES-assisted STS transfer, the HLM
reached its maximum before seat-o, and had a smaller
averaged maximum value (15%BW cm/s) than the HLM
of the healthy subjects rising with strong arm-support.
3.2.3. Reaction forces
In rising without FES, after seat-o the hands toler-
ated all dynamics of the upward movement (more than
85%BW) and thus, the VGRF decreased after seat-o to
less than 15% of patient's BW. Applying FES, at the end
of phase II the knee joints were fully stretched and the
Table 3
The average horizontal distances of the CoM from ankle and CoP under the feet at seat-o (standard deviation)
Without arm-support With normal arm-support With strong arm-support
X
ankle
CoP [cm] +1.1 (0.75) +4.15 (1.5) 3.1 (2.5)
X
CoM
CoP [cm] +4.1 (10.8) 6.6 (3.8) 7.8 (8.5)
Table 4
The averaged maximum values (standard deviation) of the joint torques and the horizontal and vertical forces at the shoulders for healthy subjects
rising with three dierent strategies
a
Without arm-support With normal arm-support With strong arm-support
Max (s
ankle
) 1.88 1.98 1.24
[%BW BH] (0.79) (0.8) (0.67)
Max (s
knee
) 5.07 4.41 2.76
[%BW BH] (0.8) (1.1) (1.29)
Max (s
hip
) 3.66 3.03 1.02
[%BW BH] (1.69) (1.28) (1.82)
Max (s
shoulder
) 4.55 10.8
[%BW BH] (0.14) (0.63)
Max (F
xshoulder
) 9.48 7.11
[%BW] (6.01) (4.63)
Max (F
y shoulder
) 6.54 23.65
[%BW] (6.01) (12.7)
a
The numerical values are calculated for one joint.
130 F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133
feet could carry up to 60% of the patient's BW. In this
case, the loads were transferred from hands to the legs
when actually 6570% of the movement duration was
passed.
3.2.4. Joint torques
In the FES-assisted standing up, the maximum total
knee joint torque was 2.0% BW BH and thus, the active
knee joint torque which was generated by the electrical
stimulation of the quadriceps was signicant. Due to the
stimulation pattern, the FES-induced knee joint torque
of the patients attained its maximum value after seat-o,
when almost 60% of the movement duration was over.
3.2.5. Shoulder loads
The vertical forces at the shoulder (Fig. 5) indicates
that the patients used their arms about and after seat-o
to compensate for the limited joint torques at the lower
extremities. In addition, after seat-o, the torque at the
shoulder has remained almost constant. This indicates
that the reaction torque at the shoulders was most
probably produced to adjust the orientation of the trunk
relative to the horizon and to stabilize the posture. It
should be noticed that the torques at the patients'
shoulder are not completely comparable with that of the
healthy subjects, since the relative positioning of the
bars and seat for healthy and paraplegic test groups
were dierent. The horizontal forces at the shoulder
might be generated to control the forwardbackward
uctuation of the body during rising. In the STS transfer
with FES at the end of phase II, the vertical forces at the
shoulder decreased (Fig. 5), which was on account of the
load sharing of the legs with the hands.
4. Discussion
According to Riley et al. [1] a number of variables
(e.g., chair height or speed of rising) eect the perfor-
mance of a STS task. However, when the task perfor-
mance is restrained, even with constraints which are
within the range of normal conditions, one must be
careful in the generalization of the results. We believe
that the STS transfer, like most of other human tasks
(e.g., reaching [27]) exhibits a number of properties
which are almost independent of the initial body state or
speed of task performance, etc. Thus, to investigate the
invariance properties of the STS task no constraint was
established on the participants in performing the de-
manded tasks.
4.1. Healthy adults
4.1.1. Stability of the movement
Using the arm-assistance, the support points of the
body and thus the stabilizing factors of the movement
were increased. Thus, the healthy test group rising with
arm-support did not try to transfer the body CoM from
the chair to the support base of the feet before leaving
the seat. However, in all cases at the seat-o instant the
CoP was under the support base of the feet (see Table 3).
In rising with arm-support the HLM at seat-o was
positive but smaller than HLM in standing up without
arm-support, whereas, in rising with normal arm-sup-
port the ascension velocity of the body CoM at the seat-
o instant was larger than in the other cases. This fact
together with considerable gravity moment arms
(X
CoM
CoP) at seat-o instant indicate that, the forces
and torques at the shoulder (or the forces and torques
applied by the hands at the bars) supported the stability
of the movement in the STS transfer with arm-assis-
tance.
4.1.2. Momentum control
Table 2 indicates that the STS task without arm-
support was accomplished more quickly than other
strategies. This fact together with the positive gravity
moment arm at the seat-o instant, suggest that in
standing up without arm-support the motion had to
be performed with more dynamics to ensure the pos-
tural stability. Since in this case the support base of
the body after seat-o is smaller than in rising with
arm-assistance. In the STS transfer with normal arm-
assistance the MTP was shorter than in the two other
strategies and as mentioned, the two peaks in the ki-
netic energy were less distinct. The horizontal and
vertical LM's at seat-o were almost equal and thus,
momentum transfer between two phases was less sig-
nicant. In rising with strong arm-support the HLM
at seat-o was greater than VLM, and the MTP was
longer than in the STS transfer with normal arm-
support. These observations along with the spatial
displacement of the body CoM suggest that in rising
with strong arm-support, the healthy participants ap-
plied a combination of the two strategies described by
Riley et al. [1] to control the upper body exion and
its velocity prior to seat-o (emphasizing more on the
strategy I).
4.1.3. Joint torque distribution
In the STS transfer with normal arm-assistance the
arms supported the stability of the movement and
thus, reduced the maximum knee and hip joint torques
which occurred shortly before seat-o instant. The
vertical reaction forces indicate that in this case the
legs compensated for the static loads as well as for a
part of the dynamic loads due to the inertial, coriolis
and centrifugal forces. Applying stronger arm-support
by the healthies, the arms not only shared with the
legs in the dynamics of the motion, but also they
carried a part of the body weight (the static loads)
which in turn resulted in the smaller joint torques at
F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133 131
the lower extremities compared with the normal arm-
support.
4.2. Paraplegic patients
We do not intend to generalize all of our results about
the paraplegic patients because, only two paraplegic
subjects participated in this experimental analysis, and
therefore no statistical data were presented. Although
the relative positions of the chair-bars were dierent for
the paraplegic and healthy test groups, nevertheless,
comparing the results with the data obtained from the
healthy participants, it is still possible to deduce quali-
tative general information.
4.2.1. Momentum control
It was noticed that in our paraplegic participants the
VLM at seat-o was almost zero and the HLM in phase
I was controlled to ensure the postural balance and to
facilitate the separation from the seat. In rising without
FES, the patients used the launching maneuver to leave
the seat, which resulted in a higher maximum HLM
than rising with FES. These observations together with
the time history of the kinetic energy of the body CoM
indicate that for the paraplegic participants the mo-
mentum transfer was negligible and the STS transfer was
accomplished in a more static manner than in the
healthier rising with strong arm-support. After leaving
the seat, the patient controlled his vertical velocity
(VLM) such that he reached the standing stance with a
zero vertical velocity while adjusting his trunk orienta-
tion. Thus, not only the position of the CoM but also its
velocity in the horizontal and vertical directions were
controlled.
4.2.2. Load distribution
The maximum total knee joint torque generated by
the stimulation of the quadriceps (2% BW BH) was
comparable with the maximum knee joint torque of the
healthy subjects rising with strong arm-support (2.76%
BW BH). Nonetheless, the vertical GRF and the LMs
indicate that applying FES, the legs participated eec-
tively in the dynamics of the movement rst when the
VLM has reached its maximum and the HLM was al-
most zero (almost 65% of the movement duration was
over).
5. Conclusion
The results of this study suggest that, locating the
support-bars in front of the healthy subjects enlarged
the support base of the body and thus, the postural
stability was increased. Applying strong arm-support,
the hip and knee joint torques were reduced by 50%. The
healthy participants rising with strong arm-support also
used momentum transfer maneuver to facilitate the
transition from sitting to standing. In the paraplegic test
group it was observed that the momentum transfer
strategy was not applied. With/without FES the HLM
was used mainly to facilitate the separation of the body
from the seat. The forward exion of the trunk was to
provide the postural balance before leaving the seat. In
the FES-induced standing up the legs could participate
partly in the movement dynamics, but the load sharing
of lower extremities was complex and most probably
subject-dependent. These observations suggest that the
healthy participants used the constant information from
the actual state of the entire body to distribute the loads
between their hands and legs, and to control the whole
body movement. For the paraplegic patients only a part
of these information was available and thus, some sig-
nicant dierences were observed between the maneuver
they have used compared with the strategy applied by
the healthy adults rising with strong arm-support.
Acknowledgements
A part of this work was accomplished during the stay
of the rst author at the Technical University of Munich
under grants of the Deutscher Akademischer Aus-
tauschdienst (DAAD) and SFB462 project Sensomotorik
sponsored by the German Research Council DFG.
Those parts of this project carried out at the University
of Tehran have been nancially supported by that uni-
versity. The authors are also indebted to Prof. Parn-
ianpour from the Ohio State University for his valuable
remarks.
References
[1] Riley PO, Schnekman ML, Mann RW, Hodge WA. Mechanics of
a constrained chair-rise. J Biomech 1991;24(1):7785.
[2] Mulder AJ, Veltink PH, Boom HBK. On/o control in FES-
induced standing up: a model study and experiments. Medical &
Biol Eng & Compt 1992;March:20512.
[3] Quintern J, Minwegen P, Mauritz KH. Control mechanisms for
restoring posture and movements in paraplegics. Progress in
Brain Research 1989;80:489502.
[4] Bajd T, Kralj A. Standing up of a healthy subject and a paraplegic
patient. J Biomech 1982;15(1):110.
[5] Donaldson N. Unsupported standing of paraplegics by stimula-
tion of plantarexors: nested linear controllers and experimental
protocol. In: Proc. EC RAFT meeting Restoration of Muscle
Activity through FES & Associated Technology, Salford, June
1993:6174.
[6] Yamaguchi GT, Zajac FE. Restoring unassisted natural gait to
paraplegics via functional neuromuscular stimulation: a computer
simulation study. IEEE Trans Biomed Eng 1990;37(9):886902.
[7] Khang G, Zajac FE, Paraplegic I. Computer model and control
system design. IEEE Trans Biomed Eng 1989;BME-36(9):87384.
[8] Donaldson NN, A topdown view of paraplegic standing. In:
Proc. of Ninth Int. Conf. on Mechanics in Medicine and Biology,
Ljubljana, Slovenia, June 30July 4 1996:4258.
132 F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133
[9] Donaldson NN, Yu CH. FES standing: control by handle
reactions of leg muscle stimulation (CHRELMS). IEEE Trans
Rehabil Eng 1996;4(4):2804.
[10] Bahrami F, Riener R, Schmidt G. Arm-supported standing up: a
comparative study. In: Proc. of Second Annual Conference of the
International Functional Stimulation Society, Burnaby, Canada,
1620 August 1997.
[11] Coghlin SS, MacFadyen BJ. Transfer strategies used to rise from
a chair in normal and low back pain subjects. Clin Biomech
1994;9(2):8592.
[12] Doorenbosch CA, Harlaar J, Roebroeck ME, Lankhorst GJ.
Two strategies of transferring from sit-to-stand: the activation of
monoarticular and biarticular muscles. J Biomech
1994;27(11):1299307.
[13] Arborelius UP, Wretenberg P, Lindberg F. The eects of armrests
and high seat heights on lower-limb joint load and muscular
activity during sitting and rising. Ergonomics 1992;35(11):1377
91.
[14] Schultz AB, Alexander NB, Ashton-Miller JA. Biomechanical
analysis of rising from a chair. J Biomech 1992;25(12):138391.
[15] Riley PO, Popat R, Krebs DE. Momentum analysis of sitback
failures in sit-to-stand trials. IEEE-EMBC and CMBEC
1995:12834.
[16] Dolan MJ, Restoration and biomechanical evaluation of stan-
ding up and sitting down in paraplegics using functional
electrical stimulation. Ph.D. Thesis. University of Strathclyde,
UK, 1996.
[17] Riley PO, Mann RW, Hodge WA. Modelling of the biomechanics
of posture and balance. J Biomech 1990;23(5):5036.
[18] Zatsiorsky V, Seluyanov V. The mass and inertia characteristics
of the main segments of the human body. In: Biomechanics VIII-
B. Baltimore, MD: University Park Press, 1983:11529.
[19] Winter DA. Biomechanics and motor control of human move-
ment. New York: Wiley, 1990.
[20] Magnan A, McFadyen BJ, St-Vincent G. Modication of the sit-
to-stand task with the addition of gait initiation. Gait & Posture
1996;4:23241.
[21] Kralj A, Jaeger RJ, Munih M. Analysis of standing up and sitting
down in humans: denitions and normative data presentation. J
Biomech 1990;23(11):112338.
[22] Hutchinson EB, Riley PO, Krebs DA. A dynamic analysis of the
joint forces and moments during rising from a chair. IEEE Trans.
Rehabil Eng 1994;2(2):4956.
[23] Hemami H. Some aspects of Euler-Newton equations of motions.
IngenieurArchiv 1982;52:16776.
[24] Craig JJ. Introduction to robotics, mechanics & control. Reading,
MA: AddisonWesley, 1986.
[25] Hof AL. Scaling gait data to body size. Gait & Posture
1996;4:2223.
[26] American Academy of Orthopedic Surgeons. Joint motion:
methods of measuring and recording. Edinburgh: Churchill Liv-
ingstone, 1966.
[27] Goodman SR, Gottlieb GL. Analysis of kinematic invariances of
multijoint reaching movement. Biol Cybernet 1995;73:31122.
F. Bahrami et al. / Clinical Biomechanics 15 (2000) 123133 133