DYNAMIC MEASUREMENT OF 3-D STIFFNESS OF ANKLE-FOOT

ORTHOSES

Paolo CAPPA, Giuseppe DI ROSA, Fabrizio PATANÈ

Paolo CAPPA (SEM member) is a Full Professor, Department of Mechanics and Aeronautics, University of Rome “La
Sapienza”, Via Eudossiana 18, 00184 Rome – ITALY. Ph.: ++39-06-44585273, Fax: ++39-06-4881759, E-mail:
[email protected] Clinical Engineering Service, Children’s Hospital ‘‘Bambino Gesù’’, Piazza San Onofrio 4,
00165 Rome, ITALY, Ph.: ++39-06-68592128, Fax. ++39-06-685924843
Giuseppe DI ROSA is a Phisician, Paediatric Neuro-rehabilitation Division, Children’s Hospital “Bambino Gesù”, Via Torre
di Palidoro – 00050 Passoscuro (Fiumicino) Rome – ITALY, Ph. ++39-06-68593421, E-mail: [email protected]
Fabrizio PATANE is a Research Associate, Department of Mechanics and Aeronautics, University of Rome “La Sapienza”,
Via Eudossiana 18, 00184 Rome – ITALY. Ph.: ++39-06-44585585, Fax: ++39-06-4881759, E-mail:
[email protected]

ABSTRACT
This paper investigates the mechanical properties of the ankle foot orthosis (AFO), i.e. the
aid routinely prescribed to control position of the ankle joint of impaired patients. However,
no procedure of objective testing of an orthosis is actually performed by the customer, i.e.
the clinician, who is unable to determine if the reaction of a patient wearing a specific AFO
proves unsatisfactory due to a non-optimized making of the device, or to an unexpected
pathological response of the subject.
The following description refers to a testing apparatus that allows a validation of the
expected mechanical performances of AFOs. Representative curves for two AFOs of the
same leg of a patient produced by one orthotist by using the same material (i.e.
polypropylene) but different in shape (i.e. one is an internal AFO and the latter is an
external one) are presented to illustrate the capabilities of the novel testing system
proposed here.

Keywords: Ankle Foot Orthosis; Stiffness; Experimental Analysis; Testing Apparatus

NOMENCLATURE
α, β, γ : dorsi/plantar flexion, inv/eversion, int/external rotation angles,
AFO: ankle-foot orthosis,
cθ sθ tθ: sine, cosine and tangent of generic angle θ,
EAFO: external ankle-foot orthosis,
  Fx   Mx 
  S  
S
f =  Fy  , m =  M y  : forces and moments measured by the load cell (at point K and
F  M 
 z  z
represented in the space S),
jP jF jM: Jacobian of positions, forces, and moments,
l: distance between K and F (see Fig.2),
IAFO: internal ankle-foot orthosis,
L: vertical component of l,
p: position of ankle joint in the fixed frame (coordinates x, y and L),
 Mα 
 
τ =  Mβ  : Anatomical moments (dorsi/plantar, inv/eversion, int/external
 
 Mγ 
moment),
uc : combined standard uncertainty.

ANATOMICAL TERMINOLOGY (See Ref.30 for more details)

Sagittal plane: the plane containing the axes of the foot and the tibia,
Frontal plane: the plane perpendicular to the axis of the tibia,
Transversal plane: the plane perpendicular to Sagittal and Frontal planes,
Plantar/dorsi flexion: the rotation lying in Sagittal plane,
Inversion/eversion: the rotation lying in Frontal plane,
Internal/external rotation: the rotation lying in Transversal plane.

1. INTRODUCTION
Orthotic devices [1] for lower limb assistance are externally applied devices generally
prescribed in patients affected by a non severe cerebral palsy to: (a) modify structural and
functional characteristics of neuromuscular and skeletal systems and (b) promote, by
exerting forces on the body, stance phase stability and enhance swing phase clearance
during walking [2-3]. The orthotic devices represent part of a physical management plan.
More specifically the solid ankle–foot orthoses (AFO) are orthotic technical aids capable of
preventing excessive plantar-flexion, which is one cause of toe walking, by correcting the
foot–shank angle in swing. Figure 1 shows two types of spiral AFOs that were analyzed in
this paper: one is of an internal type, indicated in the following as IAFO, while the latter is an
external one, named as EAFO. Both are made in polypropylene composite woven fabric and
are designed for the same leg of a patient. However, the following are repercussions of the
use of AFOs: partial loss of functional foot and ankle motion, high energy consumption and
altered proprioception [4-5]. All AFOs are built ad-hoc for each patient. The materials,
generally proposed to the clinician by the manufacturer on the basis of elastic properties
and fatigue resistance, are polymers such as polypropylene and suchlike, or, more recently,
composite materials consisting of a matrix reinforced with carbon fiberglass, aramid, or
boron fibers. The matrix can be a thermosetting resin (such as an epoxy, polyester or
polyamide) or a thermoplastic resin (such as nylon or polysulfone). Moreover, the shapes
may vary in order to obtain more or less compliance in fixed and selected directions.
In recent times, by means of the development and use of new materials and novel
geometry in AFOs [6-7] to suit individual needs, the role of AFOs has been entirely
reconsidered: they are no longer a simple foot support but they can also be considered a
complex dynamic active element that has to interfere with the biomechanics of the patient.
In fact, the main current task of AFOs is to provide a most effective stiffness evaluated in a
3-D approach in relation to the capability of controlling the movement of the ankle exhibited
by the patients.

(a) (b)

Figure 1. Spiral ankle-foot orthosis; (a) EAFO, (b) IAFO.

The flexibility of AFOs, that is recognized as the key parameter, was extensively examined
in literature both numerically and experimentally [8-20] with the additional aim of extending
the life cycle of AFOs and preventing the observed failures in clinical practice [21]. Other
papers are available in literature and are more generally focused on the experimental
evaluation of the stiffness of AFOs carried out by means of simple self-fabricated
mechanical devices that lack in an exhaustive 3-D evaluation [22-26]. The use was also
proposed of opto-electronic systems designed and commercially available for clinical gait
analysis, to conduct the flexibility measurement of AFOs [27]; however, the methodology
suffers when numerous AFOs have to be examined due to the time length necessary to
conduct each trial.
Moreover, the mechanical characteristics of an AFO are not only dependent on the chosen
geometry and material but, unfortunately, also on the manufacturing process; in actual fact,
they are custom-made from thermoplastics that are moulded over rectified plaster models
of the affected part of the body. The spread of the geometrical shapes of AFOs attained by
skilled technicians was demonstrated [28] and consequently the best approach implies, de
facto, an experimental analysis relative to each AFO prior to the current use with the
patient. In addition, due to the highly asymmetric nature of AFOs only the complete
description of the 3-D stiffness would allow the prediction of the full set of interactions that
occurs with the patient under complex in-vivo daily activities.
In order to easily overcome the previously indicated limitations in the knowledge of
mechanical behavior of AFOs the authors in a previous paper [29] reported on a novel
testing apparatus that allows: (a) the evaluation of AFO stiffness in two anatomical planes
(identified in the medical jargon as sagittal and frontal planes) and (b) the conduction of
semi-automatic trials carried out in a quasi-static manner. The global error associated to
AFO stiffness values was always less than 4 %.
Based on the results obtained with the trials conducted by means of the mentioned testing
apparatus, the present paper intends to describe the design, construction and evaluation of
an automatic tensile test system capable of overcoming the limits of previous devices
described in scientific literature. The previously indicated limits concern: range, rate and
number of displacements applied in a full 3-D approach to AFOs. Instead, the investigation
of the limit in the use of AFOs due to fatigue damage of fiber-reinforced thermoplastic
composite fabrics is not a target of the present stage of the research project. A modular
design was therefore chosen and design criteria included ease of construction with a simple
manufacturing process; the design is obtained with the aid of an industrial robot that is
capable of positional control with adequate accuracy in continuous and complex path
motions. The design uses off-the-shelf materials, components and instrumentations to
simplify manufacture and consequently reduce costs.

2. METHODS

2.1. General project description

The loading apparatus described here is similar in principle to the one described in Ref.29
and developed by the same authors of the present paper with modifications to allow
continuous cyclic loading between specified endpoints. In addition while the previously
developed apparatus was instrumented in such a manner that it could measure only two
anatomical rotations and torques, zeroing the third (i.e. the int/external moment, Mγ), the
prototype here described measures the six components of the exerted interaction force.
The basic task of the present work is to upgrade the testing machine already developed by
achieving: (a) a full automatic apparatus and (b) the 3-D measurement of the entire set of
ankle rotations and moments evaluated by dynamically testing the AFO. The feature (b)
allows a complete mechanical evaluation of the AFOs by tailoring the response of all the
three anatomical moments corresponding to any geometrical configurations imposed to the
AFO. Moreover, in respect to the semi-automatic solution proposed in Ref.29, the choice of
automating the apparatus also allows: (c) the reduction of testing time; (d) an effective
independence from operator’s skill; and, finally, (e) the conduction of dynamic trials, thus,
the viscoelastic behavior of the materials used to make the AFOs may be taken into
account.
2.2. Theoretical considerations

The joint motion schematization adopted here uses the definitions described in the ISB
proposal of the standard for joint kinematics [30]. The methodology involves the conversion
of a spherical motion in a planar one. So, it is sufficient to move point K (Figure 2) in
respect to the plantar aspect of the foot to get a deflection of the tibia around the anatomical
angles α and β, while the third and last rotation γ is simply obtained by rotating the foot
around the axis e1. The kinematics of the apparatus is reported in the following equations
that express the variations of x and z as a function of α and β with tα, cβ and tβ representing
tangent and cosine of angles α and β:
t
x = L α , z = −L t β (1)
cβ
The dynamics of the device, neglecting the inertial and gravitational components, is instead
slightly different from the results indicated in Ref.29, because the measurements of all the
components of reaction at point K are acquired in order to evaluate also Mγ, i.e. the ext/int
rotation moment, taking advantage of the full set of reaction forces made available by an
industrial 6-component load cell.

Knee

Tibia

Foot

Ankle

Figure 2. Kinematic scheme of the testing apparatus and definition of anatomical angles α, β and γ.

In particular, in the prototype described in Ref.29 the values of Mα and Mβ depend only on
Fx and Fz because Mγ is always zeroed and the friction over the prismatic joint (see Figure
2) is assumed to be unnoticeable. The general solution that has to be adopted in the
present case can be expressed as indicated in Eq.2, where the array T represents the
moments expressed as a function of anatomical planes:
 Mα   l ( s β sα Fz − c β sα Fy − cα Fx ) + c β M z − s β M y 
   
τ =  Mβ  =  l ( c β cα Fz − s β sα Fy ) + M x  (2)
   
 Mγ   l ( sα Fz + s β cα Fx ) + M y 

2.3. Experimental set-up and procedure

The apparatus consists of an industrial cartesian xy Robot (payload 100 N at the maximum
speed of 300 mm/s) with a supplementary rotary stage, mounted as indicated in Figure 3,
and equipped with 3 rotary optical encoders 2500 step/cycle. The controller is a
programmable PID, which is able to perform simple positioning tasks as linear, circular and
spline trajectory interpolations. The reaction is measured at point K by means of an
industrial 6-component load cell (error 0.1 % FS, FS=120 N). All analog and digital data are
acquired by a DAQ device (sample rate 120 Hz, resolution 14 bit).
Among the different types of orthoses one EAFO and one IAFO were examined here. The aim
of the clinician when choosing these types of orthotic devices, is to get a given stiffness
direction, or, in other words, a spring-like effect, over the ankle complex, so as to support
the patient at least during the gait phase.
The experimental procedure was fully automatic and was also chosen in order to reduce to
a unnoticeable level the effects induced by the relaxation of the material in which the AFOs
are built, as it occurred in Ref.26. So, the methodology capable of overcoming the effect of
 the creep is simply achievable by: (a)
Load cell getting the AFO to pass as many
times as possible through its un-
deformed configuration and (b)
Spiral Orthosis Universal joint applying displacements which imply
symmetrical loads, as a tendency.
Each of the two AFOs was tested five
Shank
times; the first two cycles were
Prismatic joint
considered dry runs that were carried
Spherical joint out to verify the entire set-up and the
Foot remaining cycles were instead used
for analysis. The collected data were
then averaged and expressed as total
Motor (γ)
Linear servo (x) moment Mt.
 Mβ 
 
Linear servo (z)
3
Mt (α , β ,γ ) =  cα c β Mγ  (3)
 M 
 α 

Figure 3. Testing machine.

2.4. Metrological evaluation

Due to the absence of a performance verification based on the use of test items with
calibrated stiffness value and known associated uncertainty, the metrological analysis was
conducted with a type B approach. Assuming the error sources independent and
symmetrical with regard to positive and negative values, the combined standard uncertainty
uc is:

( ) ( ) ( )
n
uc ( Mi ) = ∑j
i =1
2
pij uc2 p j + j Fij
2
uc2 F j + j Mij
2
uc2 M j (4)

Due to the kinematics characteristics of the testing apparatus, the previous equation is
unlinear. Moreover the global error mainly depends on the uncertainty associated to the
load cell due to the capability of the robot in positional control of the simulated foot. Hence,
the uncertainty of each measurement is substantially determined by the grade of flexibility
of the AFO tested. In actual fact, the stiffer the AFO, the more the load cell output
approaches its end scale and the lesser is, as a percentage, the global uncertainty.
Therefore, to avoid overestimating the uncertainty, an error map was calculated for each
AFO tested to evaluate if the global uncertainty should be evaluated as a percentage of the
full scale output or as a linear function of the total moment Mt. The results are described in
the next section.

3. RESULTS AND DISCUSSIONS

As decided in Ref.29, to synthetically represent the mechanical behavior of the AFOs,
evaluated during the experimental trials, the diagrams of the total moment Mt (see Figure 4)
were selected as a key parameter, with the difference that the present approach is more
exhaustive and representative of actual use. The comparison between the specifications
and the effective behavior of the AFOs tested gives a positive result which confirms the
flexibility direction requested by the clinical user. Indeed, it also emerges, as a general
tendency, that the isomoment lines are more spaced in β direction than in α one, which is a
consequence of the larger flexibility of AFOs in Frontal plane than in Sagittal plane. In
addition, for both the AFOs, when β is leaved constant, the isomoment lines are more
spaced as α increases, i.e. the orthoses are stiffer in Sagittal plane for plantar-flexion
angles then for dorsi-flexion rotations. Conversely, comparing the behavior of the AFOs in
Frontal plane, when α is maintained constant, the isomoment lines are more spaced as β
increases for EAFO, while the opposite happens for IAFO, which results in a stiffer behavior for
inversion movements than for eversion ones. The previously observed behavior shows that
both the orthoses qualitatively fulfil the clinical specification and the performances declared
by the manufacturer.
In the graphs of total moment Mt, the lines of maximum and minimum stiffness are also
depicted. The mentioned lines are defined by the preferential directions of flexion of the
ankle in the plane αβ, and they are identified by means of the values of angle θ. From an
examination of Figure 4, it appears that θ mainly changes, as expected, depending on
angle γ (see Table 1); in particular, when the int/external rotation is null, it gives θMAX=28°
and θMIN=16° for the EAFO and IAFO, respectively. The previous parameter, θ, could be
assumed as an index to classify the orthoses as objectively identifying the effective
flexibility direction of AFOs.
Table 1 – Angles of Maximum and Minimum stiffness measured

γ θMAX θMIN
[degree] [degree] [degree]
-15 3 5
EAFO 0 28 14
15 38 24
-15 25 24
IAFO 0 16 15
15 5 5

The type B error maps, not showed here for sake of brevity, present their maximum values
distributed near the limits of the chosen displacement field. That behavior, as observed in
section “Metrological Evaluation”, depends mainly on the metrological performances of the
load cell, following the line of maximum stiffness of the AFO tested. In addition, because the
range of variation of uncertainty proved to be of about 0.003 Nm, i.e. less than 0.02 %, it is
reasonable to assume the error as constant in all the trials performed and equal to
0.175 Nm. The uncertainty assumes, in conclusion, a maximum value of the moment equal
to 20 Nm, i.e. about 1% FSO. The indicated uncertainty is easily acceptable in relation to
the aim of the apparatus described here and represents a noticeable improvements if
compared to an error of 4% FSO exhibited by the previous version of the testing machine.

4. CONCLUSIONS
The fully automatic testing machine shows its suitability for testing the stiffness of AFOs in a
3-D manner that effectively resembles the actual use by producing continuous data for the
AFO throughout its range of motion. The uncertainty associated to the data obtained is low
enough (always less than 1%) to accurately test the AFOs. To the best of our knowledge,
there exists no data on the 3-D dynamic flexibility with which to compare the performance of
our apparatus. The machine provides, in a highly reproducible fashion, synthetic graphs of
the total moment, that can be considered a benchmark of the tested AFO.
 Mt N m (total moment) Mt N m (total moment) Mt N m (total moment)
15 15 20 15
20 16 16
20
16 12
16 20 16
16
10 θMAX 10 10 16 12
16 12 20
12 12 8
16 12 8
12 8 4
5 8 8 5 5 12
8 8
8 12 4 16
4 θMIN 8 8
4 4 4
EAFO

α ° (flex/ext)
α ° (flex/ext)

α ° (flex/ext)
0 8 0 0 4
4 4 8 12
8 12 8 12
8 12 8 16
8
12 12 16 12 16
-5 -5 12 -5 12 20
16 16 20 16
16 20 16 16
20 20 20
20 20 20
-10 -10 -10
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
β ° (inv/ev) β ° (inv/ev) β ° (inv/ev)

(a) (b) (c)

Mt N m (total moment) Mt N m (total moment) Mt N m (total moment)
15 15 15 20
20
16 20
16 16
16 16 16 16
10 12 10 10
12
12 16 12 12 12
8 12 8
16 12
5 4 5 12 5 8 8
8
4 4 8
4 4
8 4 4
8

(flex/ext)
(flex/ext)
(flex/ext)

8 12 8
IAFO

α°
α°
α°

0 4 0 12 4 0 8 4 4
12 4
16 4 12 8
16 8 16 8
20 12
8 1 20 12 8 20 12
8 24 24 16 16
12 28 16 12 28 20
-5 2824 16 -5 -5 24 20
20 16 24
20 12 24 28
20 28
16 28 24
24 28
28 20
-10 24 -10 -10
-15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15
β° (inv/ev) β° (inv/ev) β° (inv/ev)

(d) (e) (f)

Figure 4. Variation of total moment Mt as a function of α and β. The diagrams are relative to: (a) γ=-15°, (b) γ=0°, (c) γ=15°, for the EAFO; (d) γ=-15°, (e)
γ=0°, (f) γ=15°, for the IAFO. The directions of maximum and minimum stiffness are also indicated.
That value, in conjunction with the preferential directions of flexion of the ankle, represents
synthetic parameters that could be provided to the clinical user as a technical
documentation by the manufacturer. In addition, the testing apparatus described here not
only overcomes the limitations of the earlier testing device used for the objective evaluation
of the mechanical performances of AFO, but also produces a complete data set for an AFO
in about 15 minutes; hence, it makes it possible to conduct in-vitro tests for each AFO prior
the clinical use in a fully automatic fashion. The proposed graphs and benchmarks could be
valuable tools to evaluate, in an objective manner, whether or not the contribution of the
orthoses is able to achieve the treatment goals.
As a conclusive consideration it may be appropriate to note that the data made available by
the testing system might be useful not only in a clinical perspective (being objective
decisional tools for medical staff) but also in biomechanical research fields, where
experimental data are usable for musculo-skeletal simulations or dynamics calculations.

REFERENCES
1. International Organization for Standardization. (1989) ISO 8549-3: Prosthetics and
Orthotics – Vocabulary, Part 3: Terms relating to External Orthoses. ISO, Geneva.
2. L.M. Knutson., D.E. Clark “Orthotic devices for ambulation in children with cerebral
palsy and myelomeningocele” Phys Ther; 71:947–60, 1991.
3. E. Bower “A guide to physiotherapy techniques in cerebral palsy” Current Paediatrics,
9, 79–83, 1999.
4. R.B. Mollan, W.V.A James. “New flexible design of drop-foot orthosis” Injury, 8(4):
310-4, 1977
5. J.F. Lehmann., S.M., R. Condon, B.J. Price de Lateur “Gait abnormalities in
hemiplegia: their correction by ankle-foot orthoses” Archives of Physical Medicine and
Rehabilitation, 68:763-71, 1987.
6. J.Romkes, R.Brunner “Comparison of a dynamic and a hinged ankle–foot orthosis by
gait analysis in patients with hemiplegic cerebral palsy” Gait and Posture, 15:18–24,
2002.
7. A.G. Feldman, D.J. Ostry, M.F. Levin, P.L. Gribble, A.B. Mitnitski “Recent tests of the
equilibrium-point hypothesis (lambda model)” Motor-Control, 2: 189-205, 1998.
8. D. Leone, S. Diemente, S. Gustave, I. Lopez, “Structural analysis of solid ankle-foot
orthosis”, Bioengineering Conference Proceedings of the 1988 Fourteenth Annual
Northeast, pp. 26-8, 1998.
9. D. Leone, “A structural model for molded thermoplastic ankle-foot Orthosis analysis”,
ASME Journal of Biomechanical Engineering, 109, pp.305-10, 1987.
10. D. Leone, S. Diemente, M. Lopez-Isa, “Structural stability prediction for thermoplastic
ankle-foot orthosis”, Bioengineering Conference, Proceedings of the 1991 IEEE
Seventeenth Annual Northeast, pp. 231-232, 1991.
11. T. Chu, “Experimental validation on finite element stress analysis of a polymeric
orthotic device”. IEEE-EMBC and CMBEC, 2, pp. 1259-1260, 1995.
12. D. Deitz, “Optimizing orthotic designs with FEA”, Mechanical Engineering Magazine,
July issue, 119(7), pp. 70-71, 1997.
13. T.M. Chu, N.P. Reddy, J. Padovan “Three dimensional finite element stress analysis of
the polypropylene, ankle foot orthosis static analysis”, Med. Eng. Phys, 17, pp. 372-9,
1995.
14. N.P. Reddy, G. Pohit, P.C. Lam, R.C. Grotz “Finite element modelling of ankle-foot
orthoses”, Proc Inter Conf Biomechanics and Clinical Kinesiology of Hand and Foot,
Indian Institute of Technology, pp. 97-9, 1985.
15. S. Syngellakis, M.A. Arnold, H. Rassoulian “Assessment of the non-linear behavior of
plastic ankle foot orthoses by the finite element method”, Proc Inst Mech Eng, 214(5),
pp. 527-39, 2000.
16. D. Leone, S. Diemente, S. Gustave, I. Lopez “Structural analysis of solid ankle-foot
orthoses” IEEE, 26-28, 1988.
17. D. Leone “A structural model for molded thermoplastic ankle-foot Orthosis analysis”
ASME Journal of Biomechanical Engineering, 109: 305:310, 1987.
18. Deitz D. “Optimizing orthotic designs with FEA” Mechanical Engineering 1997.
19. T.M. Chu, N.P. Reddy, J. Padovan “Three dimensional finite element stress analysis of
the polypropylene, ankle foot orthosis static analysis” Med. Eng. Phys, 17: 372-379,
1995.
20. S. Syngellakis, M.A. Arnold, H. Rassoulian “Assessment of the non-linear behavior of
plastic ankle foot orthoses by the finite element method” Proc Inst Mech Eng,
214(5):527-39, 2000.
21. T. Chu “Experimental validation on finite element stress analysis of a polymeric
orthotic device” IEEE-EMBC and CMBEC, Theme 5, 1995.
22. S. Yamamoto, M. Ebina, M. Iwasaki, S. Kubo, H. Kawai, R. Kayashi “Comparative
Study of Mechanical Characteristics of Plastic AFOs” J of Prosthetics and Orthotics, 5:
259-64, 1993.
23. T.R. Lunsford, T. Ramm, J. Miller “Viscoelastic Properties of Plastic Paediatric AFOs”
J of Prosthetics and Orthotics, 6: 3-9, 1994.
24. R. Singerman, D.J. Hoy, J. Mansour “Design Changes in Ankle-Foot Orthosis Intended
to Alter Stiffness Also Alter Orthosis Kinematics” J of Prosthetics and Orthotics, 11: 48-
56, 1999.
25. Y.L. Jeng, H. Kitaoka, K. Kaufman “Orthosis Stiffness Characteristics” 6th annual
GCMA meeting, 2001.
26. B. Klasson, P. Convery, S. Raschke “Test apparatus for the measurement of the
flexibility of ankle-foot orthoses in planes other than the loaded plane” Prosthet Orthot
Int, Apr;22(1):45-53 1998 Apr;22(1):45-53, 1998.
27. K. Katdare, M. Schwartz, R. Wervey “The Non-Linear Stiffness of Ankle Foot Orthoses
Measurement and Prediction” 5th GCMA meeting Rochester Minnesota USA april 12-
15, 2000.
28. N. Sykes, J.C. Peacock, G.R. Johnson “A method for the measurement of three-
dimensional geometry of casts used for the manufacture of ankle foot orthoses”
Clinical Biomechanics, 13, 542-547, 1998.
29. P. Cappa, F. Patanè, M.M. Pierro “A novel device to evaluate the stiffness of ankle-
foot orthosis devices” ASME Journal of Biomechanical Engineering, December 125: 1-
5, 2003.
30. Letter to the editor. “ISB recommendation on definitions of joint coordinate system of
various joints for the reporting of human joint motion part I: ankle, hip, and spine”
Journal of Biomechanics, 35:543-548, 2002.