mme.modares.ac.ir
1
*2
3
- 1
- 2
- 3
* 9177948944
29 : 1394
: 17 1395
: 25 1395
. .
- -
.
"
"
. .
. -
-
. .
.
- -
.
Implementation of a trajectory predictor and an exponential sliding mode controller on a knee exoskeleton robot
Kaveh Kamali, Ali Akbar Akbari*, Alireza Akbarzadeh
Department of Mechanical Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
* P.O.B. 9177948944 Mashhad, Iran, [email protected]
A
RTICLEI
NFORMATIONA
BSTRACTOriginal Research Paper Received 19 January 2016 Accepted 06May 2016 Available Online 14 June 2016
In this article, design and hardware development of a knee exoskeleton robot is discussed. The robot aims to help the individuals with lower extremity weakness or disability during the sit-to-stand movement. In the trajectory generation phase, a new method is proposed which uses a library of sample trajectories to predict the sit-to-stand movement trajectory based on the initial sitting conditions of the user. This method utilizes the theory of "dynamic movement primitives" to estimate the sit-to-stand trajectory. The trajectory generation method is tested on a library of human motion data which has been obtained in a laboratory of motion analysis. In the next step, an exponential sliding mode controller is used to guide the robot along the predicted trajectory. The controller and the trajectory generator are implemented on the exoskeleton robot. For the hardware development, the xPC Target toolbox of MATLAB software and a data acquisition card was used. Finally, the robot was tested on a male adult.
The subjects were asked to wear the robot while doing several sit-to-stand movements from various sitting positions. According to the results, the average power which is required to be applied by the user’s knee is less when the exoskeleton assists him.
Keywords:
Exoskeleton Robot Sit-to-stand Movement Trajectory Prediction Dynamic Movement Primitives Exponential Sliding Mode Controller
- 1 . -
. -
1] .[
3-11Exoskeleton Robots
. :
-
- .
- .
-
. )
(
1-
] [
2. . .
.
-
]
.[
3. .
"
"
2.
] [
4.
( )
. -
-
. .
- -
. - -
. -
-
.
. ] [
5.
"
"
3. -
.
1 Electromyogram signals
2 Central Pattern Generator
3 Dynamic Movement Primitives
. .
. - - .
.
.
.
.
45
] [
6.
-1
- - .
- -
.
-2
. .
)
-3
] .[
5.
- - .
-4
.
:
4 Chattering
5 Exponential Sliding Mode Controller
.
.
. .
.
- 2 . )
(
1.
"
"
1. .
- 2 - 1
) (
a 2-
2
.
. .
. .
2(b)
. -
.
1
-2-2 )
( .
. .
.
1 Linear Series Elastic Actuator (LSEA)
2 Ball-Screw Fig. 1 Knee exoskeleton which is designed in FUM robotic lab
1
1 Table 1 Specifications of the LSEA
2.5 mm/rot
-
5 mm
-
50000 N/m 1000 N 125 mm/s
0. 04 mm
±
. .
.
- 3-2
] .[
8,7. )
( ]
.[
9.
. .
3 4
. .
3(a) 3(b).
3 Sagittal Plane
(a)
(b)
Fig. 2 (a) The linear series elastic actuator (LSEA) which is designed in FUM robotic lab; (b) Magnetic linear encoder in the LSEA.
)
2(
a) (
b]
[
10.
5
= 1, … ,5)
:
(1)(1)
==
==
-
. .
4 = 5 1
.
. )
(
(2)
(2) F = M(q)q + C(q, q)q + G(q)
q = [ , … , ]
.
M(q)
× 4
C(q, q)
4 ×
44
G(q) 4 × 1
F
:
(3)(3) F =
× 2× 2
× 2
.
- -
- 3
] [
5. .
) (
. - -
. .
- 1-3
]
12,11
. [
]
(4):[
11(4)
= + ( )
=
) (
y. .
> 0
)
(
= 4= 0
.
= 25
= 25 4
. -
( )]
1[
12) (
4 ( )1 Radial-basis
(a) (b)
Fig. 3 (a) Kinematic model of the human-exoskeleton system; (b) kinematic model of the knee exoskeleton
3 (a)
(b)
Magnetic tape Magnetic encoder Ball-screw Servo Motor Spring
(6,5)
:
(5)
( ) = ( )
( )
(6) ( ) = exp( ( ) )
= 1, … ,
-
x 0
> 0 1
. ) (
4.
(7)
(7)
=
= 25 3 (0) = 1
. ) (
(4) (7)
. .
] [
13.
) ( )
(
.
(8)
(8)
= + ,
= 1, … ,
.
T(10,9) (11)
(9)
F = (t )
( ) , w =
(10) X =
( ) ( )
( ) ( ) ( )
( )
( ) ( ) X × w F = 0 (11)
) (
11) )
) (
4((
7- 3 - 2 -
- .
. ] [
5. .
. -
(12)
, ,
)
,.
I Y
.
(14,13)
:
(13)
I = [ , , , , , ]
(14)
Y = [ , , , , , ]
.
ww (15)
(15) (Z): (I, Y) (w, )
.
(Z)- ) (
)
I , Y
( )
I , Y
(
. ]
:[
5(16)
= ( X w F ) × ( (Y ))
]
1[
14.
w
-
.
N k
d
(18,17)
(17) ( ) = (1 | | ) , | | < 1
0,
(18) (I , I ) = D × (I I ) , D = diag 1
, > 0 (18) (19)
(19)
= max,…, min,…, , ,
-
)
( 16
) .
( 16
(20)(20)
= ( X w F + ( ) )
× (I , I ) + (Y , Y )
) (
20)
1 Locally Weighted Regression
(12) Z = , , , , , ; (I , Y , )| = 1, … , ,
= 1, … ,
)
) ((
16(
20 (I , I ).
) (
16) .
(
20( )
) (
20)
I , Y
( .
. -
4
.
- 4 -
- :
- - .
[16,15]
.
- - .
-
.
70
-
1.72-
. .
- - .
.
2.
-
1
. -
5- .
. .
- - -
2
.
1Raptor-E Digital Real-Time Motion Analysis
2 AMTI
"
- - -
"
. -
.
[17]6
- -
- - .
-
- - .
)
(
20- - - -
. -
- -
. .
-
2-
Table 2 Experiment conditions for sit-to-stand trials
(
cm) (
cm)
-10, -7, -4, 0 ,4 40
-10, -7, -4, 0 ,4 43
-10, -7, -4, 0 ,4 46
-10, -7, -4, 0 ,4 49
Fig. 4 The procedure of the proposed trajectory generation method
4
Extracting the initial conditions , goal and the time constants of each trajectory
, , , , , | = 1, … , = 1, … ,
|
Acquiring sample points of all the training trajectories:
= , , , , , ; ( , , )| = 1, … , = 1, … , Labelling the acquired trajectories with their initial condition
vector, goal vector, and their time constant
Input the desired initial condition and goal
Minimizing the objective function to find the optimum shape parameter vector and time constant of DMP equations for the desired initial condition and goal
Using the optimum shape parameter vector and time constant in the DMP equations to generate the desired
trajectory
Fig. 5 Snapshots from the motion analysis procedure of sit-to-stand movement
5
- -
Fig. 6 Path of the shoulder, hip, knee and ankle during the sit-to-stand movement for four different seat heights
-
6-
- - .
.
(20)
= 1
= 0
(20)
(21)
= ( X w F + ( ) ) × (I , I ) (21)
- - -
. -
-
(I)
-
- -
(0) = 1
)
-
25
25 ( )
)
(
1-3 .
) (
4 (I)( )
. ) (
I) (
(21).
- -
7
-
. ) (
( )
- 5
-
- .
- -
. )
(
.
. .
8
.
) .
-
.
- )
.(
) .
- -
. ) (
8. )
.(
8- - - -
-
. .
. . .
.
. .
- 5 - 1
) (
2 .) q
4( .
) (
2(22)
= [m … m ] C(q, q)q + G(q) + 2
+ 2 + 2 +
M (q) i,j
.
- )
(
22) (
22 (23)(23)
= ( , ) + +
. .
9(25,24)
(24)
=
(25)
= + ,
sign( )
] [
18. .
sat( )
] [
18 sign( ).
. -
.
[19,6]
.
. .
[19]
-
(26)
(26)
= 1
+ + ( )
. sat( ) ( , )
sat( )
( )
(27) ( ) = + (1 )e | | (27)
]
r[
6. ) (
) .
) (
22(
23(28)
(28)
= + 2 + 2 + 2
(a)
(b)
(c)
(d)
Fig. 7 Generalization of sit-to-stand movement. Graphs (a) and (b) show the knee angle with respect to the ankle angle for different foot position and different sit heights, respectively. Graphs (c) and (d) show the knee angle with respect to time for different foot position and different sit heights, respectively
-
7- .
) ) (
a(
b.
)
) (
c(
d(28)
[m … m ] C(q, q)q + G(q)
. )
(
28.
) )
. ((
29
. .
) (
26(29)
=2 × 10 (29)
-
1
5
- 5 - 2 .
.
200±
10 V. .
.
) ) (
26(
29±
10 V.
.
.
500- -
1 2
. -
- .
( )
- -
) .(
. .
- -
. .
500
1 xPC Target
2 Matlab
. .
8 8
8
)
8.(
10.
11
- 6
. .
- -
) .(
12-
10- . -
4pred.
( )q
pred.
a pred.
a pred.
a
) , ( ) , ( )
( l l l
5 4 4,q,q q
Fb
Fig. 8 Overall diagram of trajectory estimation and control process for the knee exoskeleton
8
ˆ14
1 m
q
4q
4q
5) , (q4q4
h
pred.
a
) (l
pred.
a
) (l
pred.
a
) (l
Fb
Fig. 9 Schematic diagram of exponential sliding mode control used for controlling The knee exoskeleton
9
) (
( )
. .
) ) (
2(
. .
] [
20. -
10-
13
. )
) (
a(
b)
) (
(
.
) (
c 13- - -
- - )
(
d 13.
- -
Fig 10 Data acquisition board used for implemention of the controller 10
. .
) (
a)
14(
b.
14 1430
- - -
-
Fig. 11 Diagram of the hardware/software structure of the controller and trajectory generator of the exoskeleton
/
11Host PC Data
Acquisition Board
Ankle Rotary Encoder Foot Force Sensor
Router Magnetic Linear Encoder
Servo Motor Driver
Servo Motor of the LSEA
Force Sensor Amplifier
Fig. 12 Snapshots of the sit-to-stand movement with the help The knee exoskeleton
-
12-
]
23-21
. [
] [
23.
] [
23)
30 3
-
5- .
.
- 7
- - .
- - .
- - .
"
"
. .
. .
. -
-
. .
.
- - .
.
(a)
(b)
(c)
(d)
Fig. 13 Graphs of mean and standard deviation of: (a) knee torque; (b) knee power; (c) knee angle; (d) vertical ground reaction force, during the sit-to-stand movement
)
13) (
a) (
b(
c) (
d-
-
(a)
(b)
Fig. 14 Average power of knee for the conditions of with / without
assistance for: (a) different foot position and (b) different seat height ) : 14
( a
) (b
.
- - 30
. . . .
-
- 8
[1] A. Norhafizan, R. A. R. Ghazilla, V. Kasi, Z. Taha, B. Hamid, A review on lower-Limb exoskeleton system for sit to stand, ascending and descending staircase motion, Applied Mechanics and Materials, Vol. 541, No. 5, pp. 150-1155, 2014.
[2] N. Karavas, A. Ajoudani, N. Tsagarakis, J. Saglia, A. Bicchi, D.
Caldwell, Tele-impedance based assistive control for a compliant knee exoskeleton, Robotics and Autonomous Systems, Vol. 73, No. 1, pp. 78–90, 2015.
[3] H. K. Kwa, J. H. Noorden, M. Missel, T. Craig, J. E. Pratt, P. D.
Neuhaus, Development of the IHMC mobility assist exoskeleton. In IEEE International Conference of Robotics and Automation ICRA'09, Kobe, Japan, May 12-17, 2009.
[4] R. Ronsse, B. Koopman, N. Vitiello, T. Lenzi, S. M. M. Rossi, J.
Kieboom, E. Asseldonk, M. C. Carrozza, H. Kooij, A. J. Ijspeert, Oscillator-based walking assistance: a model-free approach, In International Conference on Rehabilitation Robotics, Zurich Science City, Switzerland, June 1-6, 2011.
[5] A. Ude, A. Gams, T. Asfour, J. Morimoto, Task-specific generalization of discrete and periodic dynamic movement primitives, IEEE Transactions on Robotics, Vol. 26, No. 5, pp. 800- 815, 2010.
[6] C. J. Fallaha, M. Saad, Sliding-mode robot control with exponential reaching law, IEEE Transactions on Industrial Electronics, Vol. 58, No. 2, pp. 600-610, 2011.
[7] A. B. Schultz, N. Alexander, J. Ashton-Miller, Biomechanical analyses of rising from a chair, Journal of Biomechanics, Vol. 25, No. 12, pp.1383–1391, 1992.
[8] N. B. Alexander, A. B. Schultz, D. N. Warwick, Rising from a chair:
effects of age and functional ability on performance biomechanics, Journal of Gerontology, Vol. 46, No. 3, pp. 91–98, 1991.
[9] J. Music, R. Kamnik, V. Zanchi, M. Munih, Human body model based inertial measurement of sit-to-stand motion kinematics, WSEAS Transaction on Systems, Vol. 7, No. 3, pp. 156-164, 2008.
[10] F. Amirouche, Fundamentals of multibody dynamics: theory and applications, pp. 225-393, Boston: Birkhäuser, 2006.
[11] A. J. Ijspeert, J. Nakanishi, S. Schaal, Learning rhythmic movements by demonstration using nonlinear oscillators, IEEE/RSJ International Conference on Intelligent Robots and Systems, Lausanne, Switzerland, Sep. 30-Oct. 4, 2002.
[12] A. J. Ijspeert, J. Nakanishi, S. Schaal, Movement imitation with nonlinear dynamical systems in humanoid robots, IEEE International Conference on Robotics and Automation, Washington DC, US, May 11-15, 2002.
[13] S. Schaal, P. Mohajerian, A. Ijspeert, Dynamics systems vs. optimal control—A unifying view, Progress in Brain Research, Vol. 165, No.
1, pp. 425-445, 2007.
[14] S. Schaal, C. G. Atkeson, Constructive incremental learning from only local information, Neural Computation, Vol. 10, No. 8, pp.
2047-2084, 1998.
[15] R. B. Shepherd, H. P. Koh, Some biomechanical consequences of varying foot placement in sit-to-stand in young women, Scandinavian Journal of Rehabilitation Medicine, Vol. 28, No. 2, pp.
79-88, 1996.
[16] M. M. Khemlani, J. H. Carr, W. J. Crosbie, Muscle synergies and joint linkages in sit-to-stand under two initial foot positions, Clinical Biomechanics, Vol. 14, No. 4, pp. 236-246, 1999.
[17] M. W. Whittle, Clinical gait analysis: A review, Human Movement Science, Vol. 15, No. 3, pp. 369-387, 1996.
[18] J. J. E. Slotine, W. Li, Applied Nonlinear Control, pp. 276-306, New Jersey: Prentice-Hall, 1991.
[19] M.H.Rahman, M. Saad, J.P. Kenné, P.S. Archambault, Control of an exoskeleton robot arm with sliding mode exponential reaching law, International Journal of Control, Automation and Systems, Vol.
11, No. 1, pp.92-104, 2013.
[20] NASA-STD-3000 Man-Systems Integration Standards, Volume I, Section 3, Houston: NASA Johnson Space Center, 1995.
[21] S. Mefoued, S. Mohammed, Y. Amirat, G. Fried, Sit-to-stand movement assistance using an actuated knee joint orthosis, 4th IEEE RAS & EMBS International Conference on Biomedical Robotics and Biomechatronics (BioRob), Rome,Italy, June 24-28, 2012.
[22] N. Karavas, A. Ajoudani, N. Tsagarakis, J. Saglia, A. Bicchi, D.
Caldwell, Tele-impedance based stiffness and motion augmentation for a knee exoskeleton device, IEEE International Conference on Robotics and Automation (ICRA), Karlsruhe, Germany, May 6- 10, 2013.
[23] A. Gams, T. Petric, T. Debevec, J. Babic, Effects of robotic knee exoskeleton on human energy expenditure, IEEE Transactions on Biomedical Engineering, Vol. 60, No. 6, pp. 1636-1644, 2013.
