RAGHU INSTITUTE OF TECHNOLOGY

Affiliated to JNTUK, Approved by AICTE

Accredited by NBA
Ranked A Grade by AP State Knowledge Mission
Ranked 10th among top 20 private emerging engineering colleges in
India by higher education review magazine.
www.raghuinstech.com

DEPARTMENT OF MECHANICAL ENGINEERING

III B.Tech II Semester

ROBOTICS

UNIT IV
 Day 29

Topics covered: Introduction to differential transformation

Differential transformation

The trajectory planning problem is generally formulated in tool configuration space. However
the control over the motion of the robot tool is indirect, because we can controls only the speed
on rate at which each joint moves. The tool configuration specifies the position of the tool tip and
orientation of the tool. The relationship between the joint variable  and the tool configuration
is specified by the kinematics equations. If we use x to denote the tool configuration vector,
then the direct kinematics equation can be formulated in terms of the tool configuration function
f and joint variable  as follows: x = f()

Given a trajectory x(t) in tool configuration space, on way to find a corresponding trajectory (t)
in joint space is to invert the above equation and there by solve the inverse kinematics problem
directly. However, finding an explicit closed form solution to the inverse kinematics problem can
be a difficult task, depending upon the kinematic complexity of the manipulator.

Another method for controlling the joints of a manipulator in a coordinate’s position is Jacobian
control. This control method is based on the differential relationship between ‘  ‘ and ‘ x ‘.
Essentially, the Jacobian allows the computation of a differential change in the tool coordinate
frame due to a differential change, in the position of the manipulator’s joints. By making use of
the inverse Jacobian, the differential changes in the manipulator joints needed to accomplish a
differential change in the tools coordinates frame can also be computed.

Jacobian.

Fig 4.1

If xi is the coordinate of the tool point (location given by x , y , z coordinates reference to the
manipulator’s base and rotations about the x , y and z -axes associated with the tool), then the
generalized function relating each coordinate of the tool to the joint variables n can be written
as xi = fi(1……….n)

Where the maximum value of j corresponding to the number of Cartesian coordinates

and n corresponds to the number of joints of the particular manipulator. Taking the total
differential of above equation

xi = ∑
( ( ,………..

xi = ∑

Where Jij the matrix obtained from the derivatives of f with respect to  called the “tool
configuration Jacobian matrix”, the “tool Jacobian matrix” or simply the “Jacobian”. The tool
Jacobian matrix J is a linear transformation which maps the instantaneous joint space velocity 
into the instantaneous tool configuration velocity x as shown in the figure. If the matrix is
invariable, then it is possible to calculate the joint velocities given the velocities associated with
the displacement and orientation of the tool point.

Fig 4.2

The Jacobian is a time varying quantity since it must be evaluated for instantaneous joint
velocities and the actual position of the joints. It is clearly dependent on the kinematic structure
of the arm and the coordinate system used to express the tool coordinate frame. For a six degrees
of freedom robot the above differential relation may be written as:
The 6x6 coefficient matrix relating the tool velocities to the joint velocities is the “Jacobian
matrix”.

Home work:

1.What is meant by differential transformation? Explain .

2.Write the Jacobian matrix.

Previous JNTUK Questions

1.What is meant by Jacobian. 4m April 2017 & April 2016

2.What is the purpose of differentil transformation. 4m April 2015

Day 30

Topics covered : Problems

1. Two arm planar robot:

 Fig 4.3

From the direct kinematics, we have

x = a1cos1 + a2 cos(1+2)

y = a1sin1 + a2 sin(1+2)

let, cos = C , sin = S

x = a1C1 + a2 C12

y = a1S1 + a2 S12

By differentiating these two equations with respect to 1 and 2 respectively then

= − − = −( +

= −

=( +

The above two equations can be written as

x=− − (1 + 2

y= 1 + (1 + 2
 $ −( + −
#% & = ' () *
+

X=J

Where J – jacobian matrix

Now the joint velocities can be found from the end-effectors velocities via the inverse Jacobian,
i.e.

 = J-1X

Where,

−( + −
J=+ ,=' (
+

' (
−( + −( +
J-1 = - - .

The control technique described by the above equation is known as “resolved motion rate
control”. Here the motion of the robot in tool-configuration space is resolved into its joint space
components through the tool configuration Jacobian matrix.

2. Three arm planar robot:

Fig 4.4

From the direct kinematics, we have

X = a1c1+a2c12+a3c123

Y = a1s1+a2s12+a3c123
α = 1 + 2 + 3

By differentiating these two equations with respect to and 1 2 respectively then

The above two equations can be written as

Where the matrix „ J ‟ is the Jacobian matrix. Now the joint velocities can be found from the
end-effector‟s velocities via the inverse Jacobian, i.e.

 = J-1 X

Where

Day 31

Topics covered : Singularities, robot dynamics.

Singularities
The linear transformation relating the joint velocities to Cartesian velocities is given by

 = J-1 X

Where, J-1 is the inverse Jacobian matrix. One of the potential problems solving for the joint-
space velocity in this manner is that at certain points in joint space, the tool Jacobian matrix loses
rank, i.e. there is reduction in the number of linearly independent rows and columns. As we
approach these points in space J() becomes “ ill conditioned ”, with numerical solution of
above equation producing very large joint velocities. The points at which J() looses its rank are
called “joint-space singularities”.

Mathematically this can be defined as follows: consider a n link manipulator and denote its
Jacobian matrix as J() . Let an integer ' n is the maximum rank of J(). If a configuration  = 
s satisfies, rank of J()< n then we denote this s as the singular configuration.

By examine the Jacobian on its inverse, singularities of the work space may be identified. The
presence of entries that can make the determinant zero indicate the nonexistence of a
corresponding inverse and can be used to define positions or areas in the work space where the
use of a Jacobian and its inverse are illegal. We should also examine the conditions that have a
tendency to make the determinant as zero. For these cases while the condition that forces the
determinant to be zero may not be physically attainable by the manipulator, velocities of some
joints may reach. Large values which may be beyond the capabilities of the control system.
These joint space singularities are classified into two types.

Boundary Singularities : These occur when the manipulator is fully stretched on or folded back
on itself such that the end effector is near or at the boundary of the work space.

Interior Singularities : These occur away from the work space boundary and generally are
caused when two or more of the axes of the robot form a straight line, i.e. become collinear.
Boundary singularities are not particularly serious, because they can be avoided by simply
performing the manipulations sufficiently far from the surface of the work envelop. However, in
the case of interior singularities, the effects of a rotation about one of the axes can be cancelled
by a counter acting rotation about the other axis.

Robot dynamics

Ans:

During the work cycle, a manipulator must accelerate, move at constant speed, and decelerate.
This time-varying position and orientation of the manipulator is termed as its dynamic behavior.
To balance the internal and external forces, time-varying torques are applied at the joints by joint
actuators. The internal forces includes, inertial, coriolis and frictional forces are caused by
motion of links. The external forces includes, load and gravitational forces are the forces exerted
by the environment. As a result, links and joints have to withstand stresses caused by forces or
toques balance across these. In this, mathematical model for the dynamic behavior of the
manipulator is developed. The mathematical equations, often referred as manipulator dynamics,
are a set of equations of motion (EOM) that describes the manipulator dynamics response to the
input actuator torques. The dynamic model of a manipulator is useful for computations of torque
and force required for execution of a typical work cycle. This is the vital information for the
design of links, joints, drives and actuators. This behavior of the manipulator provides the
relationship between the joint actuator torques and motion of links for simulation and design of
control algorithms. This control maintains the dynamic response of the manipulator to obtain the
desired performance which depends on the accuracy of dynamic model and efficiency of control
algorithms. To achieve the desired response and performance, the control problem requires
specifying the control strategies. This simulation permits testing of control strategies, motion
planning and performance studies without a physical prototype of the manipulator. Generally the
complex dynamic system can be modeled systematically by using the physical laws of
Lagrangian mechanics and Newtonian mechanics. The approaches such as Lagrange – Euler
(LE) which is energy based and Newton – Euler (NE) based on force balance can be used
systematically for developing the manipulators equations of motion (EOM). These EOM are a
set of second order, coupled, non-linear differential equations consists the inertia loading and
coupling reactions between the joints. The LE and NE formulations of a model provide a closed
form solution which is computationally intensive making real time control. These two
formulation methods provide a symbolic solution to manipulator dynamics and insight into the
control problem.

Home work:

1.Explain about robot dynamics

2.Explain about singularities

Previous JNTUK Questions:

1.Determine the manipulator jacobian matrix and singularities for the 3-DOF articulated arm.

16m April 2017

2. What are the singularities of a manipulator? How are they classified? 8m April 2016

Day 32 & 33

Topics covered: Lagrangian mechanics

Lagrangian mechanics with an example

Lagrangian function or Lagrangian ‘ L ’ is a scalar function, defined as the difference between

the total kinetic energy (K.E) and the total potential energy (P.E) of the manipulator,
i.e. L =K.E – P.E

The LE dynamic formulation is based on a set of generalized coordinates to describe the system
variables. In generalized coordinates, the generalized displacement ‘ q ‘ is used as a joint
variable, in which ‘ d ‘ is used for prismatic joint and ‘  ‘ for angular displacement of a rotary
joint. Then the linear and angular velocities are represented by ‘ d ‘ and ‘  ‘ respectively for
prismatic and rotary joints. The torques required to produce desired dynamics, ‘ f ‘ represents the
force for prismatic joints and ‘  ‘ for rotary joints. Therefore, the Lagrangian ‘ L’ is the function
of K.E and P.E which are the functions of qj and qj. , where 1,2,............. n

The dynamic model based on the Lagrangian – Euler formulation is obtained from the
Lagrangian as a set of equations:

From the above equation .

# &= Torque due to kinetic energy

/ 1
/0 2

# & = Force due to potential energy

1
2

j = Joint torque for a joint which can be provided by actuator.

If j = 0 then joint does not move.

If j  0 then, the manipulator movement is modified by the actuator at joint ‘ j ‘

Example: 2 DOF manipulator dynamic models

Generally, to illustrate the Lagrange – Euler formulation and to clarify the problems in the
dynamic modeling, the dynamics of a simple manipulator is the working part. Let us consider a
planar, 2 – DOF manipulator with both rotary joints as shown in the below figure.
 Fig 4.5

The dynamic model can be obtained from the general direct geometric approach as follows: for
this model,

Let 0 and 1 are the coordinate frames,

1 and 2 = Joint variables

m1 and m2 = Mass of links which are assumed as point masses that located at the centre of mass
for each link and the links are slender members.

L1 and L2 = Link lengths

V1 and V2 = linear velocities

1 and 2 = Angular velocities

The Lagrangian is the function of kinetic and potential energies of the manipulator. Therefore,
the kinetic energy of a rigid body can be expressed as

K.E = mv2 + I2

Where v = Linear velocity

 = Angular velocity

I = mass moment of inertia of rigid body at its centre

The Lagrange – Euler formulation for link 1 is
Where

Mii = Effective Inertia

Mij = Effective Coupling Inertia

Hi = Centrifugal Force

Gi = Coriolis Acceleration forces

When the manipulator is more than 2 – DOF then computation of the equations will be
cumbersome as the numbers of equations are increased. Therefore for obtaining the EOM for an
n – DOF manipulator, we will use homogeneous transformation matrices.

Home work:

1.Determine the equations of motion for 2DOF RR- planar manipulator arm using Lagrange-
Euler Formulation.

Previous JNTUK Questions:

1.Determine the equations of motion for 2DOF RR- planar manipulator arm using Lagrange-
Euler Formulation. 16m April 2016

Day 34 & 35

Lagrange – Euler formulation

It is a systematic procedure for obtaining the EOM and dynamic model of an n – DOF
manipulator. Let us consider an n – DOF open kinematic chain serial link manipulator which has
„n‟ joint positions or displacement variables q = [ q1 q2 q3 …………..qn ]T

According to Lagrange – Euler formulation

gives the relationship between the joint positions, velocities, and accelerations and also
generalized torques applied to the manipulator.

The derivation of EOM using LE formulation can be carried out according to the following sub-
steps. In this we will use link transformation matrices T,

 First, the link velocity is computed.

 The link inertia tensor will be evaluated.
 By using above two steps, the kinetic energy of the link can be computed.
 After obtaining the potential energy, the dynamic model of the n – DOF
  manipulator can be computed by using the equations obtained in the previous three steps.

Step 1: Velocity of a point on the manipulator

To compute the kinetic energy of a link of an n – DOF manipulator, the link velocity can be
computed as follows: Consider a point P on a link n of a manipulator as shown in the
following figure. As per the conventions the coordinate frames are chosen as frame {0}, frame
{n-1} and frame {n}. The position vector nrn describes the point P on the link with respect to
frame {n}. Then
n
rn = 3$ % 4 156

The position of point P with respect to base coordinate frame is

0
rn = 0Tn nrn = ( 0T1 1T2 …………n-1Tn ) nrn

Fig 4.6
Step 2: The inertia tensor

During the motion of the link, the mass contributes inertia forces, which reflect all the inertial
loads with respect to rotations about the origin of frame of interest, are represented by a moment
of inertia tensor. This was defined as follows:
Step 3: Kinetic energy and Potential energy

The kinetic energy and potential energy of the manipulator are calculated by using the following
relations

Step 4: Equations of Motion

Home work:

1.Determine the equations of motion for 2DOF RR- planar manipulator arm using Lagrange-
Euler Formulation.

Previous JNTUK Questions:

1.Determine the equations of motion for 2DOF RR- planar manipulator arm using Lagrange-
Euler Formulation. 16 m April 2017

2.What is Lagrange – Euler formulations? What are its applications? 4m April 2015

Day 36 & 37

Characteristics of Lagrange – Euler formulation

 It is systematic and describes motion in real physical terms.

 The equations of motion obtained are analytical and compact.
 The matrix – vector form of equations is appealing for calculations and control system
design.
 The control problem can be simplified by designing the structure of the manipulator with
minimal joint couplings, i. e. coefficients Mnj and hnjk may be reduced or eliminated.
 The model is computationally intensive and is not amenable to online control.

Newton Euler formulation.

This method of dynamic modeling of robotic manipulators is based on the Newton’s second law
and D’Alembert principle. The balance of all forces acting on a link of the manipulator leads to a
set of equations, whose structure allows recursive solution. A forward recursion describes the
kinematic relations of a moving coordinate frame, is performed for propagating velocities and
accelerations followed by a backward recursion for propagating forces and moments. The initial
position, velocity and acceleration of each joint are assumed to be known. The joint torques
required causing these time – dependent motions to realize a trajectory are computed using the
recursive NE dynamic equations of motion.

Step 1: Recursive Newton – Euler formulation

 Fig 4.7

In this formulation the serial open kinematic chain structure of a manipulator is exploited. The
NE formulation requires two passes over the links of the manipulator, one for computing the
velocities and accelerations of the links, and second, to compute the joint forces and torques as
shown in the above figure. The forward iteration or outward iteration is carried out to compute
the velocities and accelerations of each link recursively, starting from the base and propagating
forward to the end-effector. The boundary conditions are base velocity, linear and angular, which
are zero, and the boundary acceleration is the gravitational acceleration. In the backward or
inward iteration the forces and the moments acting on the links are computed using Newton’s
and Euler’s equations. First, the force and moment applied at the centre of each link must be
computed and then the force and moment applied at each joint are computed starting from the
end-effector and moving inward to the base. The end-effector force fn+1 and moment n+1
exerted by the environment or grasped object provide the necessary boundary conditions.
Forward iteration: The equations for velocity and acceleration with respect to the base frame
{0} are given as follows: The linear velocity of link i with respect to reference frame is
As the manipulator moves, the parameters referred to the base frame will be changed which in
turn the computations are complex. The computations are simplified by referring all velocities,
accelerations, inertia tensors and location of centre of mass of each link to their own link-
coordinate frames. Hence,
Backward iteration:

After finding the velocities and accelerations of each link, the forces and moments are computed
by starting at the end-effector and working backwards. Let i f and i be the force and moment
exerted by link (i-1) on ith link at the origin of frame {i-1} with respect to base frame {0} as
shown in the following figure. The total external force Fi and moment Ni acting on ith link are
computed as follows;

Fig 4.8
Home work:

1.Explain about Newton – Euler formulations by considering an example.

Previous JNTUK Questions:

1.Explain about Newton – Euler formulations by considering an example. 16m April 2017

2. What is Newton – Euler formulations? What are its applications? 4m April 2016

3. What is Lagrange – Euler formulations? What are its applications? 4m April 2015

Day 38

Topics covered: Differences between lagrangian and newton formulation

Difference between Lagrangian and Newton formulation

Home work:

1.Differentiate between Lagrange Euler and Newton Euler Formulation.

Previous JNTUK Questions:

1.Differentiate between Lagrange Euler and Newton Euler Formulation.4m April 2016