ME5402/EE5106 Advanced

Robotics (AY21-22 Semester 2)

ME5408 Kinematics of Robot
Manipulators (AY21-22 Mini
Semester 2A)
CHUI Chee Kong
Control & Mechatronics Group
Mechanical Engineering, NUS
About Me…
 PhD, Biomedical Precision Engineering Lab, The
University of Tokyo, Japan, 2004
 MSc (By Research), BSc (Hon), National
University of Singapore, Singapore
 Research interests: Medical devices and
robotics, imaging and simulation
 Website: http://blog.nus.edu.sg/mpecck/
 Email: [email protected]

2
Course Mechanics
 Lecture: Monday (6:00pm-9:00pm) via Zoom
 Class info, lecture slides, notes and tutorials can
be found in LumiNUS

3
Main Text
 John J Craig, “Introduction to Robotics:
Mechanics and Control”, Pearson, 2014, ISBN
10:1-292-04004, ISBN 13:978-1-292-04004-2
 Haruhiko Asada and Jean-Jacques E Slotine,
“Robot Analysis and Control”, John Wiley and
Sons, 1986, ISBN:0-471-83029-1
 Other references will be announced in class or
via LumiNUS.

4
ME5402/EE5106

Continuous Assessment (CA) – 70%

Part I (35%)
 Online Quiz (10%) – Recess Week
 Computing Project (25%) – Due on Week 13; a group
of at most 3 students
 Model a robot manipulator, kinematics analysis,
computer simulation of the robot manipulation
Part II (35%) – To be announced
Final Exam – 30%
 Two questions (each with multiple parts) for Parts I
and II respectively
5
ME5408

Continuous Assessment (CA) – 100%

 Online Quiz (30%) – Recess Week
 Mini Computing Project (70%) – Due on Recess Week;
individual

6
Chapter 0 –
Introduction to
Robotics
CHUI Chee Kong, PhD
Control & Mechatronics Group
Mechanical Engineering, NUS
What is a robot?
Originally from Czech language, robota, which
means forced labour
The word Robot was introduced to the public by the
Czech writer Karel Čapek in his play R.U.R.
(Rossum's Universal Robots), published in 1920.

 Dictionary.com
 1. a machine that resembles a human and
does mechanical, routine tasks on command.
 2. any machine or mechanical device that
operates automatically with humanlike skill.
8
Contents
1. History and Applications
2. Mechanics: Kinematics and Dynamics
3. Design
4. Trajectory Generation
5. Control
6. Programming

9
1. History and Applications

 Middle of 20th century

 Research in artificial intelligence (AI) – connection between
human intelligence and machines
 First robot
 1948 – William Grey Walter invented robots Elmer and Elsie that mimic
lifelike behavior using simple electronics
 1954 – George Devol invented the first digitally operated and a
programmable robot called the Unimate.
 1956 – Devol and his partner Joseph Engelberger formed the world's first
robot company.
 1961 – First industrial robot, Unimate, went online in a General Motors
automobile factory in USA.
 Advances in mechanics, controls, computers and electronics
 Robotics: The science & technology of robots

10
History (continue)

 1960s:
 Numerical control machines for precise manufacturing
 Teleoperators for remote radioactive material handling

 Late 1970s:
 Industrial robots became essential components in the
automation of flexible manufacturing systems
 1980s:
 Robotics: defined as the science which studies the
intelligent connection between perception and action

11
History (continue)

 1990s:
 Field robotics to address human safety in hazardous
environments
 Human augmentation
 Service robotics
 2000 and beyond:
 Human-centered and life-like robotics

Reference: Bellis, Mary. "Who Invented Robots?" ThoughtCo, Oct. 16, 2017,
thoughtco.com/timeline-of-robots-1992363.

12
 Robots and Robotics
Robotics is the engineering science and technology of robots, and Robotics is the study
their design, manufacture, application, and structural disposition. of system science
that involves
External World
- Perception
- Cognition
- Action
Actuators Mechanisms

Power
supply

Embedded Sensors
Control

The Honda Humanoid

Robot ASIMO
(http://world.honda.co
m/ASIMO)
Human user(s)
Applications: Manufacturing

Material handling system

Assembly line (Vecco Integrated
Automation, Inc)

14
Applications: Construction

Construction robots developed by the Takenaka Company: A crane

robot is used to place steel reinforcing bars (left), another
distributes concrete (right) (Takenama Komuten Co Ltd)

15
Applications: Farming

Sheep-shearing robot (University

of Western Australia)

Citrus-picking robot (concept)

16
Applications: Farming (continue)

The prototype robot is part of the Robot-Assisted

Precision Irrigation Delivery (RAPID) system developed
by researchers from UC Berkeley, UC Davis and UC
Merced, USA.

IEEE The Institute: Special

issue on agtech, April 2018,
http://theinstitute.ieee.org/static/
special-report-agtech

17
Applications: Military service and other
hazardous work

The US Army’s Autonomous Land Vehicle

A prototype of the Explosive

Ordinance Disposal robot (OAO
Corporation)

18
Applications: Underwater/Oil and gas

ROV: remotely
operated underwater
vehicle, is a tethered
underwater vehicle.

ROV working on a subsea structure

19
Applications: Service industry

Robot street cleaner

Robot usher Security guard

20
Applications: Entertainment

21
Applications: Education

Lego mindstorm NXT

22
Applications: Aiding the handicapped and
the elderly

Robot aids walking

Autonomous wheelchair

23
Applications: Robot-assisted surgeries
Zeus

ORTHODOC

daVinci
24
From Industrial Needs to Medical Robotics

Need for Industrial Robotics

machine
Industrial Automation
to assist
and/or
replace
worker

(www.kuka-
robotics.com)
Clinical
Service Robotics Medical Robotics Applications
Medical robotics includes
a number of devices used
for surgery, medical
training, rehabilitation
therapy, prosthetics, and
assisting people with
(www.mobilerobots.com) (www.davincisurgery.com) disabilities.
 This module focuses on mechanics and
control of mechanical manipulator

Classical Robotics == the study of

industrial robotic manipulator, which is an
integrated approach of
 Kinematics: study of geometry of
motion
 Statics: study of a manipulator at
equilibrium
 Dynamics: study of causes of motion
 Trajectory planning: generating the
path the robot must trace
 Control strategy: executing the path
 Physical hardware: building the robot

26
2. Mechanics: Kinematics and Dynamics

Kinematics is the study of position and derivatives of position without

regard to forces which cause the motion.

Dynamics concerns with the motion of bodies under the action of

force.

27
2.1 Kinematics
 Kinematics: Study of motion without regard to forces/torques that
cause it
 Interested in position, velocities, accelerations, etc. of each joint and
link
 Forward and inverse kinematics: Relationship between “joint’s
position” and “position and orientation of end-effector”

forward
Joint Space Cartesian
kinematics
Space (End-
inverse effector)
kinematics

28
Position and Orientation Representation

 How to locate objects (e.g links of manipulator,

Parts, Tools, etc) in three-dimensional space
 Frame: a coordinate system rigidly attached to each
object
 How to describe position and orientation of one frame
with respect to another frame

29
Differential (Instantaneous) Kinematics

 Jacobian of manipulator, J
 Mapping from velocities in joint space to
velocities in Cartesian space
y

2
J
1 x

 Singularities: When mapping not invertible

30
Differential (Instantaneous) Kinematics

 Jacobian also used to map static force in

Cartesian space to joint torques in joint
space
fy
2
JT
1 fx

31
2.2 Dynamics

 Equations of Motion of the Robotic

Manipulator: Describe forces required to
cause motion
Joint Forces Joint Motion
and/or Torques
Motion of Each Link

  C(q, q )q  g (q)  τ

M (q)q

Inertia Centrifugal and Gravity Joint

32 torques
Coriolis
 2.3 Robot System Components

internal sensory feedback

(position, velocity)

Robot Controller Actuator Drivers Robotic

(computer) (amplifiers) Manipulator

Task Planner
Task Planning -
Robotic task can be broken
down into motion
commands, force/torque
Desired Robotic Task
commands and strategies to
33 react to forces/torques.
3. Design
 Type of joints
 Actuators and power transmission
 Degrees of freedom
 Specialized vs universal (min 6
joints)
 Dexterity Considerations
(Geometry, Workspace)
 Speed, size, load capability
 Rigid vs Flexible
 Sensors

34
 positioning/orienting

Robotic Tasks
force/moment exerted on
environment

An Open Kinematic Chain, Serially Connected

end-effector
joints: power giving, performing
connecting mechanisms robotic tasks
35
Robot Joints

Rotational Translational
Accuracy Non-Uniform Consistent
Kinematics Complex Simple
Control Coupled, Difficult Decoupled, Easy
Link Design Simple Complex
Dexterity Good Bad

As the number of rotational joints increase,

• task planning complexity
• control algorithm complexity
• dexterity
• accuracy

36
Types of Robot Joints

Two basic types:

link j
Rotational/Revolute/
Rotary
link i joint axis

link i link j Translational/Prismatic/

Linear

37
Joints
 Each consists
of an actuator
(e.g. motor),
mechanical
transmission,
physical
structure,
sensors, etc

Physical structure typically consists of two

rigid pieces constrained to rotate or linearly
move wrt each other. Bearing is needed to
reduce the movement friction.
German Aerospace Center (DLR)
Light Weight Robot (LWR)
38
Actuator Technologies

Source of power to drive joints:

Pneumatic: Electric Motor: Hydraulic:
- energy efficient - clean - can deliver large
- hard for feedback - choice of today forces
control - bulky, leakage
problems

Note: Air-activated tools have built-in compliance important

when manipulating objects to prevent damage
End-effectors: often are pneumatic tools

39
 Mechanical transmission: takes actuator motion and direct it to the
joint
 Enable change of rotational direction
 Enable change of axis
 Torque multiplication (or reduction)
 Speed reduction (or multiplication)
 Convert rotary motion to linear motion
 Provide a “match” between the actuator & load in order that the
maximum energy is transferred to the load
 Ideal gears
 Assumptions:
 Gears are perfectly round

 Rotae on their true centers

 Inertialess

 Frictionless  No loss

 Rigid shaft attached to the gears

40
Degrees-of-Freedom (DOF)
3 for position

Rigid body in 3D Space  6 DOF 3 for orientation

Z
YAW

ROLL
PITCH
Y

41
 Degrees-of-Freedom (DOF)
In robotics/mechanism,
DOF = number of independent position variables that would
have to be specified to locate all parts of the (rigid-body)
mechanism
E.g. Planar four-bar linkage only one DOF
(even though having three moving members)

Fixe
d
42
Degrees-of-Freedom

In robotics,
DOF = number of independently driven joints

As DOF positioning accuracy

computational complexity
cost
flexibility
power transmission is more difficult

43
 Classification by Coordinate Systems
Cartesian Cylindrical

Spherical

Cartesian: x, y, z (three linear coordinates)

Cylindrical: z, r, theta (two linear and one
angular coordinates)
Spherical: r, and two angles (one linear
and two angular coordinates)

44
 Workspace is the locus of positions and
Workspace orientations achievable by the end-effector
of a manipulator.

Dextrous Workspace Reachable Workspace

“locus of tool positions for which the “locus of tool positions for which the
tool can be oriented in all possible tool can reach regardless of its
ways” orientation”

Dextrous workspace is usually much smaller

than reachable workspace
45
Operating Environment
 Clean room robots
 evacuated internally with suction in order to scavenge particles
generated by friction surfaces
 use special non-shedding materials and employ magnetic
washers to hold ferromagnetic lubricants in place
 Harsh environments (e.g. spray painting)
 clothed in a shroud in order to minimize the contamination of its
joints by the airborne paint particles

Courtesy of AIST, Japan

46
Performance/Specifications
 Resolution
 Control resolution: Smallest incremental change in tool
position that (servo) control system can distinguish (assume
no deadband, sensor errors, computational problems)
 depend on types of joints, resolution of joint position
sensors, number of joints, etc.

One-dimensional illustration of resolution

Equally spaced points control resolution

where end point can be
commanded to go

47
 Performance/Specifications
 Resolution (cont)
 If include effects of mechanical inaccuracies and sensor
errors, there is a zone about the ideal point where it may
stop.
 Spatial resolution: worst-case distance between two
adjacent positions.

Variation due to control resolution

Spatial resolution
mechanical
inaccuracies and
sensor errors
48
Performance/Specifications
 Accuracy
 measure of the ability to place the tool tip at an
arbitrarily prescribed location in the workspace
One-dimensional illustration of accuracy

Planned
closest
Target
position position
(control
resolution)
49
Performance/Specifications
 Accuracy
 measure of the ability to place the tool tip at an arbitrarily
prescribed location in the workspace
= worst-case distance from target position
= ½ x Spatial resolution

Planned
closest
Target
position position
(control
resolution) 50
Performance/Specifications
 Accuracy
 measure of the ability to place the tool tip at an arbitrarily
prescribed location in the workspace

Global vs Local accuracy

Considering In
entire neighborhood
workspace of specified
points

51
Performance/Specifications
 Repeatability
 ability of a manipulator to reposition its tool tip at a
position to which it was previously commanded.
 important for repetitive tasks

One-dimensional illustration of repeatability

Second Initial
try position

52
Maximum Speed

tool tip speed

92 mm/sec 9,000 mm/sec
Adept One SCARA
Westinghouse Series 4000

300 mm
Cycle Time : 25 mm 25 mm
Adept “Stroke”
a b
For AdeptOne
Payload Cycle Time
1 lb 0.9 sec
13 lbs 1.3 sec
20 lbs 1.7 sec

53
4. Trajectory Generation
 Each joint is prescribed with a smooth function of
time
 Coordinated motion of joints to provide desired
end-effector motion
Trajectory is a time
based function which
specifies the position
(and higher
derivatives) of the
robot mechanism for
any value of time.

54
5. Control

 Motion (position) Control

 Stable and robust algorithm to coordinate joint motion
and enable the robot to follow a specified trajectory
 Point-to-Point

 Trajectory Following

 Independent Joint Control

 Inverse Dynamics Control (Computed
Torque)
 Nonlinear control approach
 Robust control
 Adaptive control
55
Control
 Force Control
 Ability of manipulator to control forces of contact
 Complementary to position control
 Hybrid position/force control, e.g. window washing task

 Compliance control
 Impedance control

56
Motion Control Methods
 Point to Point
 Path between points not explicitly controlled
 Applications
 spot welding

 pick and place

 loading and unloading

 Continuous Path
 end-effector must follow a prescribed path in 3D space
 speed may vary or may need to be accurately
controlled
 Applications
 spray painting

 arc welding

 gluing
57
 6. Programming
 Robot programming:
 Lead-through programming
 Textual robot languages
 Lead-through programming: Desired movements are stored in controller
memory
 Powered lead-through: Using teach pendant
 Pros: Simple for point-to-point movements

 Cons: Tedious for complex movements

 Key applications: spot welding, material handling, etc.

 Manual lead-through (also called walk-through): Robot’s end-effector

moved physically by programmer, mainly used for continuous path
movements
 Cons: Difficult for large robot arm

 Key applications: Spray painting, arc welding, etc.

58
 Lead-through programming:

59
 Textual robot languages:
Task description (In plain English)
 1979: VAL, a common
textual robot language
developed by Stanford
University
 1984: VAL II (Updated
version of VAL)
 Other textual robot languages:
 Manufacturing control VAL
language (MCL)
 RAIL
 Automatic programming
tooling (APT)
 AML & AUTOPASS by IBM
for assembly operations
 Etc.
 How could desired motions be
easily implemented on robot? 60
EXAMPLE: SCARA
• Selective Compliance
Assembly Robot Arm
• Jointed cylindrical
manipulator
• Joints 1 and 2 are not gravity
loaded
• Joints are direct-drive motors
End-effector compliance
can be controlled to a certain
extent (Useful for assembly
operations which require
pegs or screw insertions

61
62
Classical robotics == the study
of industrial robotic manipulator
An integrated approach of
 Kinematics: study of geometry of motion
 Statics: study of a manipulator at equilibrium
 Dynamics: study of causes of motion
 Trajectory planning: generating the path the
robot must trace
 Control strategy: executing the path
 Physical hardware: building the robot

63
Homework
 Linear Algebra Review
 Matlab Total Academic Headcount license
 https://nusit.nus.edu.sg/services/software_and_os/sof
tware/software-student/

64
2021 Year of Robots?
https://youtu.be/BFK9lkez32E

65