Dynamic Modelling Of An Elephant Trunk Like

Flexible Bionic Manipulator

Department of Mechanical Engineering,

Indian Institute of Technology Kharagpur,
India

Presented By:
Mrunal Kanti Mishra
IIT Kharagpur, India

International Mechanical Engineering Congress and Exposition

IMECE2019
November 11-14, 2019, Salt Lake City, UT, USA
2 Authors

 M. K. Mishra, IIT Kharagpur, India

 A. K. Samantaray, IIT Kharagpur, India
 G. Chakraborty, IIT Kharagpur, India
 A. Jain, IIT Kharagpur, India
 P. M. Pathak, IIT Roorkee, India
 R. Merzouki, CRIStAL, Université de Lille, France

Mrunal Kanti Mishra 13/11/2019

3 Outline
Introduction

System Description

Forward Kinematics

Dynamic Analysis

Adams Modelling

Workspace

Conclusion

Mrunal Kanti Mishra 13/11/2019

 Introduction
4

Rigid robot Serpentine robot [1] Soft robot [2]

(Source:-https://www.renderhub.com) (Source:-Ishiguro Laboratory) (Source:-Festo)

DoF= 3-8 DoF~ 12-20 DoF = infinite

Advantages of discrete Robot [2] Advantages of soft robot

• Speed • Lightweight structure
• Accuracy • High dexterity
• High load capacity • High safety
Disadvantages of discrete robot Disadvantages of soft robot
• Low dexterity • Less accurate
• Low safety • Low speed
• Low controllability
Mrunal Kanti Mishra 13/11/2019
 System Description
5

Tube-3 Tube-2 Tube-6 Tube-5

y
z Tube-1 Tube-4

A z

Tube– 3,6
A
Base y
Base of Section-1 Section-2
Manipulator Tube– 1,4 x

1200
Flexible trunk
Tube– 2,5

Mrunal Kanti Mishra 13/11/2019

6
Forward Kinematics
By Continuum Curvature (CC) model [3] , [4]:
Homogeneous
Actuator Configuration Transformation Action
 3  li 3  li 2  
States States Matrix Variables i  tan 1  ,
 li 2  li 3  2li1 
 
lij ki   ,  ,  Ti  4*4 i 5*1
fexplicit ftypical feclectic 2 li21  li22  li23  li1li 2  li 2li3  li3li1
i  ,
 li1  li 2  li3  * rbi
Actuator Configuration Action Zone Action
li1  li 2  li 3
Zone Zone Zone i  .
3
Forward Kinematics

𝑐𝛽𝑖 1 − 𝑐𝛼𝑖
𝑐𝛽𝑖 𝑐𝛼𝑖 −𝑠𝛽𝑖 𝑐𝛽𝑖 𝑠𝛼𝑖
𝜅𝑖
𝑠𝛽𝑖 1 − 𝑐𝛼𝑖
𝑖ℎ 𝐑 𝑧 𝛽𝑖 0 𝐑 𝑦 𝛼𝑖 𝐏𝑖 𝑖ℎ
= 𝑠𝛽𝑖 𝑐𝛼𝑖 𝑐𝛽𝑖 𝑠𝛽𝑖 𝑠𝛼𝑖
𝐓𝑖𝑏 = 𝐓𝑖𝑏
𝜅𝑖
0 1 0 1 𝑠𝛼𝑖
−𝑠𝛼𝑖 0 𝑐𝛼𝑖
𝜅𝑖
0 0 0 1

Mrunal Kanti Mishra 13/11/2019

7 Forward Kinematics
2ℎ 1𝑏 1ℎ 2𝑏 2ℎ
Transformation matrix for the trunk: 𝐓𝑏𝑎𝑠𝑒 = 𝐓𝑏𝑎𝑠𝑒 𝐓1𝑏 𝐓1ℎ 𝐓2𝑏

𝐑 𝐏
[4×4]
0 1

Position (form P): Rotation (form R):

𝑖ℎ
𝑖ℎ
𝑥𝑖 = 𝐓𝑏𝑎𝑠𝑒 1,4 −𝐓𝑏𝑎𝑠𝑒 2,3
𝜃𝑥𝑖 = 𝑡𝑎𝑛2−1 𝑖ℎ
𝑖ℎ 𝐓𝑏𝑎𝑠𝑒 3,3
𝑦𝑖 = 𝐓𝑏𝑎𝑠𝑒 2,4 𝑖ℎ
−1
−𝐓𝑏𝑎𝑠𝑒 1,3
𝑖ℎ 𝜃𝑦𝑖 = 𝑡𝑎𝑛2
𝑧𝑖 = 𝐓𝑏𝑎𝑠𝑒 3,4 𝑖ℎ
𝐓𝑏𝑎𝑠𝑒 3,3

Mrunal Kanti Mishra 13/11/2019

8 Dynamic Analysis

Using Euler-Lagrange’s energy method [5]:

d K K P
 
dt  i  i  i
 Fi where
1 n
2 i 1

Ki   mi basevih
T base
vih  baseih
T base
Jih baseih 
Pi  mi g T basehi  Mi

M  C , N  ,     p

Mrunal Kanti Mishra 13/11/2019

9 ADAMS Modelling:
Fi1,act

zi  1
rb,i 1 Fi 3,act
Spring-damper system Actuation force
Fi 2,act
mi
z
li 1 Section i ,head
Fg
li 3 li 2 Bushing
Fi1, s Fkt ,i
Fct ,i Fi 3, s
Fi 2, s
x

Fi1,d
Fi 3,d Plate
Fi 2,d

Fi1,act Fi 3,act Fi 2,act

zi
Manipulator base
rb,i
Gravity
o Section i ,base

Mrunal Kanti Mishra 13/11/2019

 Workspace Of The Manipulator
Material parameters for each section
10 Parameters Values 0.34 0.34

z axis (in m)

z axis (in m)
Plate thickness 0.005 m 0.27 0.27
klinear 5300 N/m
0.2 0.2
clinear 8 Ns/m -0.08 0 0.08 -0.1 0 0.1
x axis (in m) x axis (in m)
ktorsion 0.4 N/m/rad (a) (b)

ctorsion 0.5 Ns/rad 0.1

klinear, bushing 1000 N/m 0

clinear, bushing 3 Ns/m
-0.1
-0.08 0 0.08
x axis (in m)
Trunk dimension (c)
Parameters Values (in cm)
Base diameter1 5.37
0.32
Tip diameter1 4.64

z axis (in m)
0.3
0.28
Base diameter2 4.64
Volume: 0.26
Tip diameter2 3.91 0.24
659.926 cm3 0.22
length1 10.14 0 0.05
0.05 0 -0.05 -0.05
length2 10.44
Mrunal Kanti Mishra 13/11/2019
11 Case study 1 : Rotation about negative y - axis
Force (in Case- Case- Case

(in m)
N) 1 2 -3 0.135 0.35
x

end-effector position
F11,act 0 300 350

bellow tube length (l)

backbone length ( )
z
F12,act 300 0 50

(in m)
F13,act 300 0 200 0.1175 0.225
F21,act 0 300 300
2
F22,act 350 0 75 1
l11 l12 , l13 l21 l22 , l23
F23,act 350 0 250 0.1 0.1
0 1.75 3.5 0 1.75 3.5
time (in s) time (in s)
Fi1

Fi2 Fi3
0.2 0

Tip rotation (in degree)

Vz Ry

velocity (in m/s)

rb Vx
0 -10
y
rb
0
rb
-20
600

x -0.2 -30
0 1.75 3.5 0 1.75 3.5
√3𝑟 𝐹𝑖2 + 𝐹𝑖3 time (in s) time (in s)
𝑀 = 𝐹𝑖3 − 𝐹𝑖2 𝑖 + 𝐹𝑖1 − 𝑟𝑗
2 2
Mrunal Kanti Mishra 13/11/2019
12 Case study 2 : Rotation about positive y - axis
Force (in Case- Case- Case

(in m)
N) 1 2 -3 0.24
0.125
F11,act 0 300 350

end-effector position
l11 l21 1 2

bellow tube length (l)

backbone length ( )
F12,act 300 0 50 x
F13,act 300 0 200

(in m)
z
0.1125 0.205
F21,act 0 300 300
F22,act 350 0 75
F23,act 350 0 250 l22 , l23 l12 , l13
0.1 0.17
0 1.75 3.5 0 1.75 3.5
Fi1
time (in s) time (in s)

Fi2 Fi3 28

Tip rotation (in degree)

Ry
rb 18
0 y
rb rb
8
600

x -2
0 1.75 3.5
√3𝑟 𝐹𝑖2 + 𝐹𝑖3 time (in s)
𝑀 = 𝐹𝑖3 − 𝐹𝑖2 𝑖 + 𝐹𝑖1 − 𝑟𝑗
2 2
Mrunal Kanti Mishra 13/11/2019
13 Case study 3 : Rotation about arbitrary axis
Force (in Case- Case- Case

(in m)
N) 1 2 -3 0.13 0.25

end-effector position
l21 l23
F11,act 0 300 350 0.2

bellow tube length (l)

backbone length ( )
F12,act 300 0 50 0.15 x

(in m)
F13,act 300 0 200 0.115 0.1 y
F21,act 0 300 300 0.05 z
F22,act 350 0 75 l12 l22 2 1 l13
l11 0
F23,act 350 0 250 -0.05
0.1
0 1.75 3.5 0 1.75 3.5
time (in s)
time (in s)
Fi1

Fi2 Fi3 20

Tip rotation (in degree)

Rx
Ry
rb Rz
0 y
0
rb rb
600

x -20
0 1.75 3.5
√3𝑟 𝐹𝑖2 + 𝐹𝑖3 time (in s)
𝑀 = 𝐹𝑖3 − 𝐹𝑖2 𝑖 + 𝐹𝑖1 − 𝑟𝑗
2 2
Mrunal Kanti Mishra 13/11/2019
14 Conclusion

 The model of the manipulator is developed in ADAMS MBD software. This

model simplifies the dynamical study and provides more flexibility to
introduce closed-loop control laws for actuating the manipulator.
 The manipulator tip can reach any coordinate within the manipulator’s
workspace by using two axis bending for two sections.
 The flexibility of the manipulator causes speed reduction in movement of
the tip from one position to other. Here, the manipulator takes some finite
time to settle down to its steady state.
 the path control of the tip is more difficult than the rigid link manipulator. So,
effective control scheme is to be explored in future for this kind of
manipulator.

Mrunal Kanti Mishra 13/11/2019

15 References

1. Wang, Z., Gao, Q., and Zhao, H., “CPG-inspired locomotion control for a snake robot
basing on nonlinear oscillators.” Journal of Intelligent & Robotic Systems, Vol. 85 No. 2
(2017): pp. 209-227.
2. Trivedi, D., Rahn, C. D., Kier, W. M., and Walker, I. D. “Soft Robotics: Biological Inspiration,
State of the Art, and Future Research.” Applied Bionics and Biomechanics, Vol. 5 No. 3
(2008): pp. 99-117.
3. Mahl, T., Hildebrandt, A., and Sawodny, O., “A variable curvature continuum kinematics for
kinematic control of the bionic handling assistant.” IEEE transactions on robotics, Vol. 30
No.4 (2014): pp. 935-949.
4. Webster III, R. J., and Jones, B. A., “Design and kinematic modeling of constant curvature
continuum robots: A review.” The International Journal of Robotics Research, Vol. 29 No. 13
(2010): pp. 1661-1683.
5. Falkenhahn, V., Mahl, T., Hildebrandt, A., Neumann, R., and Sawodny, O., “Dynamic
modeling of bellows-actuated continuum robots using the Euler–Lagrange formalism.” IEEE
Transactions on Robotics, Vol. 31 No. 6 (2015): pp. 1483-1496.

Mrunal Kanti Mishra 13/11/2019

16

Mrunal Kanti Mishra 13/11/2019