PICK AND PLACE TASK USING SCARA

BY SINGHMAR HARBAKSH SINGH

OUR ROBOT

 SCARA- selective compliance articulated robot arm

 Consists of three joints . 2R and 1P
 Task is to pick a ball and drop at another location

Fig. SCARA robot

PROCEDURE FOR SIMULATING THE ROBOT

Path Pt 2 Path Pt 3

1. Choose the desired path points and velocity on these path points
in cartesian space.
2. Use Inverse kinematics to convert it into joint coordinates.
3. Use Jacobian to convert it into joint velocities.
Path Pt 1 Path Pt 4
4. Generate a continuous trajectory through these joint coordinates.
5. Use forward kinematics for simulating the robot.
SIMPLIFIED DIAGRAM OF ROBOT AT ZERO POSITION

𝜃2
𝜃1 L3 𝜃2
L2 𝑑3 L2 L3 {2} 𝑑3
L4 {1} L4
L1
L1
𝜃1 -end effector frame
End effector
x-axis
-base frame y-axis
Base z-axis
HOMOGENEOUS TRANSFORMATION MATRIX

[ ]
0 0
𝑅1 𝑃1
𝑇 =¿
0 1

𝜃1

[ ][ ] [ ]
c os ⁡( 𝜃1 ) −sin ⁡(𝜃 1) 0 −1 0 0 𝑎 ∗ c os ⁡( 𝜃 1 )
a {1} 0
𝑅 =¿ sin ⁡(𝜃 1)
1 cos ⁡( 𝜃1) 0 0 1 0
0
𝑃 =¿ 𝑎 ∗ sin ⁡(𝜃1 )
1
{0} 0 0 1 0 0 −1 0

At zero position

[ ]
− c os ⁡(𝜃 1 ) − sin ⁡( 𝜃1 ) 0 𝑎 ∗ cos ⁡(𝜃 1 )
x-axis −sin ⁡( 𝜃 1) cos ⁡( 𝜃1 ) 0 𝑎 ∗ sin ⁡( 𝜃 1 )
y-axis 𝑇 =¿
z-axis 0 0 −1 0
0 0 0 1
FORWARD KINEMATICS

[ ]
𝑐 𝑜𝑠( 𝜃1) − sin ⁡(𝜃1 ) 0 𝐿 2∗ cos ⁡(𝜃 1)

𝜃2
𝑇 10 ¿ sin ⁡(𝜃1 )
0
cos ⁡(𝜃 1)
0
0
1
𝐿 2∗ sin ⁡(𝜃 1)
𝐿1
L2 L3 {2} 𝑑3
0 0 0 1

{1}

[ ]
L4 𝑐 𝑜𝑠( 𝜃2) −sin ⁡(𝜃 2) 0 𝐿 3 ∗ cos ⁡( 𝜃2 )
L1 𝑇 1
2 ¿ sin ⁡(𝜃 2)
0
cos ⁡(𝜃 2)
0
0
1
𝐿 3 ∗ sin ⁡( 𝜃2 )
0
𝜃1 {3} 0 0 0 1

[ ]
x-axis
{0} y-axis
z-axis
1 0 0 0
𝑇 23
¿ 0
0
1
0
0
1
0
− 𝐿4 +𝑑3
0 0 0 1
FORWARD KINEMATICS

𝜃2
L2 L3 {2} 𝑑3
{1} L4
L1
{3}
𝜃1
x-axis
{0} y-axis
z-axis

[ ]
𝑐 𝑜𝑠 ( 𝜃 1+ 𝜃 2) − sin ⁡( 𝜃 1 +𝜃 2) 0 𝐿 3 ∗ cos ( 𝜃 1+ 𝜃 2 ) + 𝐿 2∗ cos ⁡( 𝜃 1)

¿ sin ⁡(𝜃 1 + 𝜃 2 )
0
cos ⁡( 𝜃 1 +𝜃 2)
0
0
1
𝐿 3 ∗ sin ( 𝜃 1+ 𝜃 2 ) + 𝐿 2∗ sin ⁡(𝜃 1 )
𝑑 3+ 𝐿 1− 𝐿 4
0 0 0 1
INVERSE KINEMATICS

 Find joint variables corresponding to end effector position (, ).

[ ] ¿[ ]
.. .. .. 𝑥1 𝑐 𝑜𝑠 ( 𝜃 1+ 𝜃 2) − sin ⁡( 𝜃1 +𝜃 2) 0 𝐿 3 ∗ cos ( 𝜃 1+ 𝜃 2 ) + 𝐿 2∗ cos ⁡( 𝜃 1)
𝑇 03=¿ .. .. .. 𝑦1 sin ⁡(𝜃 1 + 𝜃 2 ) cos ⁡( 𝜃 1 +𝜃 2) 0 𝐿 3 ∗ sin ( 𝜃 1+ 𝜃 2 ) + 𝐿 2∗ sin ⁡(𝜃 1 )
.. .. .. 𝑧1
0 0 1 𝑑 3+ 𝐿 1− 𝐿 4
0 0 0 1 0 0 0 1

 Comparing the position elements

squaring and adding T(1,4) and T(2,4)

2* * *cos()=
=
INVERSE KINEMATICS

 From T(1,4) =

(* + )*cos() – * =
Using entity, A*cos()+B*sin()=C *cos()- *sin()=
 We will get

()
where = ()
JACOBIAN MATRIX

[ ]
𝐽 11 𝐽 12 .. .. 𝐽 1 𝑛
 It’s a function of joint variables which when multiplied with joint velocities gives the end effector velocity.
𝐽 21 𝐽 22 .. .. 𝐽 2 𝑛
` = J( ,…., )* = 𝐽 31 𝐽 32 .. .. 𝐽 3𝑛
𝐽 41 𝐽 42 .. .. 𝐽 4𝑛
𝐽 51 𝐽 52 .. .. 𝐽 5𝑛
End effector 𝐽 61 𝐽 62 .. .. 𝐽 6𝑛

𝜃˙ 1 𝜃˙ 1 𝜃 2=120
𝜃 2=0
DERIVATION OF JACOBIAN MATRIX

𝑙𝑖𝑛𝑘 𝑖  For Revolute Joint

𝑧 𝑖− 1
𝜔 𝑒𝑖 = 𝜔𝑖 = 𝐷 𝑖 −1 𝜃˙ 𝑖 𝑤h𝑒𝑟𝑒 𝐷 𝑖 −1= 𝑅𝑖 −1 ^
0
𝑃 𝑖 −1 𝑧
𝑒

𝜃𝑖
𝑃 0𝑒
𝑃 0𝑖 −1
𝐽𝑒𝑖=
[ ( 𝐷 𝑖 −1 𝑋 𝑃 𝑖𝑒−1 )
𝐷𝑖 − 1 ]
 For Prismatic Joint

𝑉 𝑒𝑖 =𝑉 𝑖 = 𝐷 𝑖 −and ˙
1 𝑑𝑖

𝑑𝑖
𝐽𝑒𝑖=
[ ( 𝐷𝑖 −1 )
0 ]
𝑧𝑖 − 1 𝑙𝑖𝑛𝑘 𝑖
ROBOT

𝜃2
𝜃1 L3 𝜃2
L2 𝑑3 L2 L3 {2} 𝑑3
L4 {1} L4
L1
L1
𝜃1 -end effector frame
End effector
x-axis
-base frame y-axis
Base z-axis
JACOBIAN MATRIX

= =
=

=
JACOBIAN MATRIX …

=
=

= -

=
JACOBIAN MATRIX …

==

𝜃2
L3
𝜃1 L2 𝑑3
L4

L1
End effector

Base
JACOBIAN MATRIX

𝑅𝑎𝑛𝑘=3 , 𝑠𝑜𝑤𝑒 𝑐𝑎𝑛 𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑡h𝑒 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦 𝑖𝑛3 𝑙𝑖𝑛𝑒𝑎𝑟 𝑑𝑖𝑟𝑒𝑐𝑡𝑖𝑜𝑛𝑠 𝑏𝑢𝑡 𝑛𝑜𝑡 𝑎𝑛𝑔𝑢𝑙𝑎𝑟 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑖𝑒𝑠 .

= inv()*
RECAP

1. Choose the desired path points and velocity on these path points
in cartesian space.
2. Use Inverse kinematics to convert it into joint coordinates.
3. Use Jacobian to convert it into joint velocities.
4. Generate a continuous trajectory through these joint coordinates.
5. Use forward kinematics for simulating the robot.
VELOCITY

 velocity is a free vector

THANK YOU