Differential Kinematics

• Relations between motion (velocity) in joint space and

motion (linear/angular velocity) in task space (e.g.,
Cartesian space)

• Instantaneous velocity mappings can be obtained

through time derivation of the direct kinematics or in a
geometric way, directly at the differential level

• Different treatments arise for rotational quantities

 establish the link between angular velocity and time
derivative of a rotation matrix
 establish the link between angular velocity and time
derivative of the angles in a minimal representation of
orientation 1
Differential Kinematics: the Jacobian matrix
In robotics it is of interest to define, besides
the mapping between the joint and
workspace position and orientation (i.e. the
kinematic equations), also:
 The relationship between the joints and
end-effector velocities:

 The relationship between the force applied

on the environment by the manipulator and
the corresponding joint torques

These two relationships are based on a linear operator, a matrix J,

called the Jacobian of the manipulator. 2
Differential Kinematics: the Jacobian matrix
   
x  q1 
 
y  
  q2 
w z  Cartesian Velocity  
   q3 
 x q  Joint Velocity
 y   q4 

w Jq
 
 z  
 q5 
 
q 
 6
If it 'exists' we can define the Inverse Jacobian as:

q  J 1 w

 The Jacobian is a mapping tool that relates Cartesian

velocities (of the n frame) to the movement of the
individual robot joints
 The Jacobian collectively represents the sensitivities of
individual end-effector coordinates to individual joint
displacements 3
 The Jacobian matrix

In robotics, the Jacobian is used for several purposes:

 To define the relationship between joint and workspace
velocities
 To define the relationship between forces/torques between
the spaces
 To study the singular configurations
 To define numerical procedures for the solution of the IK
problem
 To study the manipulability properties

4
 (Angular velocity of a rigid body)
“rigidity” constraint on distances among points: ||rij|| = constant

vPi − vPj orthogonal to rij

vP2 − vP1 = ω1  r12 (A)
vP3 − vP1 = ω1  r13 (B)
vP3 − vP2 = ω2  r23 (C)
P1, P2, P3: (B)−(A) = (C) 
ω1 = ω2 = ω
vPj  vPi    rij  vPi  S ( ) rij  rij    rij
• the angular velocity ω is associated to the whole body (not to a point)
• if  P1, P2 with vP1 = vP2 = 0: pure rotation (circular motion of all Pj  line
P1P2)
• ω = 0: pure translation (all points have the same velocity vP)

5
 Velocity domain
• The translational and rotational velocities are considered separately
• Let us consider two frames:
 F0 (base frame) and
 F1 (integral with the rigid body)

•The translational velocity of point p of the rigid body, with respect

to F0, is defined as the derivative (w.r.t time) of p, denoted as p:
dp
p
dt
6
 Velocity domain
For the rotational velocity, two different definitions are possible:
 A triplet  
3
giving the orientation of F1 with respect to F0 (Euler,
RPY,... angles) is adopted, and its derivative is used to define the
rotational velocity  :
d

dt
 An angular velocity vector ω is defined, giving the rotational velocity of
a third frame F2 with origin coincident with F0 and axes parallel to F1

The velocity vector ω is placed in

the origin, and its direction
coincides with the instantaneous
rotation axis of the rigid body

7
Jacobian: Analytical and Geometrical expressions
• The two descriptions lead to different results concerning the
expression of the Jacobian matrix, in particular in the part
relative to the rotational velocity
• One obtains (respectively) the:
 Analytic Jacobian JA
The end-effector pose is expressed with reference to a minimal
representation in the operational space; then, we can compute the
Jacobian matrix via differentiation of the direct kinematics function w.r.t.
the joint variables
 Geometric Jacobian JG
The relationship between the joint velocities and the corresponding end-
effector linear and angular velocity

These two expressions are different (in general)!

8
 Two problems

Problem 1: Integration of the rotational velocity ω

  dt   (orientation of the rigid body)

  dt  ??

9
10
 The integration of ω does not have a clear physical interpretation
11
So γ is the winner? NO!
Problem 2: while ω represents the velocity components about the
three axes of F0, the elements of  are defined with respect to a frame
that:
a) is not Cartesian (its axes are not orthogonal to each other)
b) varies in time according to γ

12
 Problem 2
• v and ω are “vectors”, namely are elements of vector spaces
o they can be obtained as the sum of single contributions (in any order)
o these contributions will be those of the joint velocities

• On the other hand, γ (and dγ/dt) is not an element of a vector

space
o a minimal representation of a sequence of two rotations is not obtained
summing the corresponding minimal representations (accordingly, for their time
derivatives)

 In general ω ≠ dγ/dt

However, the two expressions physically define the

same phenomenon (velocity of the manipulator) and
therefore a relationship between them must exist.
13
Finite and infinitesimal translations

Finite Δx, Δy, Δz or infinitesimal dx, dy, dz translations

(linear displacements) always commute

14
 Finite rotations do not commute

We just saw an example:

However…

15
Infinitesimal rotations do commute!
Infinitesimal rotations dφX, dφY, dφZ around x, y, z axes

16
 In summary
The two expressions of the Jacobian matrix physically define
the same phenomenon (velocity of the manipulator) and
therefore a relationship between them must exist
For example, if the Euler angles φ, θ, ψ are used for the
triplet γ, it is possible to show that

Note that matrix T(γ) is singular when sin θ = 0.

In this case, some rotational velocities may be
expressed by ω and not by  , e.g.
ω = [cosφ, sinφ, 0]T

These cases are called representation

singularities of γ.

17
Definition of matrix T(γ):

18
If sinθ = 0, then the components perpendicular to z of the
velocity expressed by  are linearly dependent x 
2
  2
y   
2
,
while physically this constraint may not exist!
From:

one obtains:

19
 Finally...:
In general, given a triplet of angles γ, a transformation matrix
T(γ) exists such that

Once the matrix T(γ) is known, it is possible to relate the

analytical and geometrical expressions of the Jacobian matrix:

Then

20
 Until now:
• We saw how we can define velocities in a
robot/rigid-body environment

• We know the connection between the analytical

Jacobian and the geometric Jacobian

• Now we calculate both of them

21
 Analytical Jacobian
The analytical expression of the Jacobian is obtained by
differentiating a vector x = f(q) ∈ 6, that defines the position
and orientation (according to some convention) of the
manipulator in F0
By differentiating f(q), one obtains

where the m × n matrix

is called the Jacobian matrix or JACOBIAN of the manipulator

22
 Analytical Jacobian
If the infinitesimal period of time dt is considered, one obtains

that is

relating the velocity vector expressed in F0 and the joint

velocity vector

 The elements Ji,j of the Jacobian are nonlinear functions of

q(t): therefore these elements change in time according to the
value of the joint variables
 The Jacobian’s dimensions depend on the number n of joints
and on the dimension m of the considered operative space:

23
AJ-Example: 2 DOF manipulator

24
AJ-Example: 2 DOF manipulator

25
Geometric
Jacobian

26
Geometric Expression of the Jacobian
• The geometric expression of the Jacobian is obtained considering
the rotational velocity vector ω
• Each column of the Jacobian matrix defines the effect of the i-th
joint on the end-effector velocity and it is divided in two terms
• The first term considers the effect of on the linear velocity v,
while the second one on the rotational velocity ω, i.e.

• Therefore

 The analytic and geometric Jacobian differ for the rotational part
 In order to obtain the geometric Jacobian, a general method
based on the geometrical structure of the manipulator is adopted
27
 Derivative of a Rotation Matrix
• Let’s consider a rotation matrix R = R(t) and R (t ) R T
(t )  I

• Let’s derive the equation: R (t ) R (t )  I  R (t ) R (t )  R (t ) R (t )  0

T T T

• A 3 × 3 (skew-symmetric) matrix S(t) is obtained

S(t )  R (t )R (t ) T

• As a matter of fact

• Then
R (t )  S(t ) R (t )
• This means that the derivative of a rotation matrix is expressed as
a function of the matrix itself

28
 Derivative of a Rotation Matrix
Physical interpretation:
Matrix S(t) is expressed as a function of a vector
representing the angular velocity of R(t)

Consider a constant vector p′ and the vector p(t) = R(t)p′

The time derivative of p(t) is

p(t )  R (t )p
which can be written as

(This last result is well known from the classical mechanics of rigid
bodies)

29
 Derivative of a Rotation Matrix
• Moreover it can be shown that:

i.e. the matrix form of S(ω) in a frame rotated by R is the

same as the skew-symmetric matrix S(R ω) of the vector ω
rotated by R

• (1) Note also that S(ω) is linear in its argument:

S(k1ω1+k2ω2 )=k1S(ω1 )+k2S(ω2 )

• (2) Note also the property of S(ω):

S(ω) p  ω  p
30
 Derivative of a Rotation Matrix
Consider two frames F and F’, which differ by the rotation R (ω′ = R ω)

Then S(ω′) operates on vectors in F’ and S(ω) on vectors in F

Consider a vector va’ in F’ and assume that some operations must be
performed on that vector in F (then using S)
It is necessary to:
1. Transform the vector(s) from F’ to F (by RT)
2. Use S(ω)
3. Transform back the result to F’ (by R)
That is:

equivalent to the transformation using S(ω)

31
Example

32
 Geometric Jacobian
The end-effector velocity is a linear composition of the
joint velocities

Each column of the Jacobian

matrix defines the effect of
the i-th joint on the end-
effector velocity

33
Geometric Jacobian

( ≜ )

34
 Geometric Jacobian – Link Velocity
Linear velocity
(of link i as a function of
velocities of link i−1)

vi−1,i denotes
the velocity of the
origin of Frame i
with respect to the
origin of Frame i−1

35
 Geometric Jacobian – Link Velocity
Angular velocity
(of link i as a function of velocities of link i−1)

36
 Geometric Jacobian – Link Velocity

Prismatic joint:

Revolute joint:

37
 Geometric Jacobian – Computation n
v n
v  qi  J Pi
qi
i 1 qi i 1
Linear velocity
Joint i prismatic

Joint i revolute

38
 Geometric Jacobian – Computation

Angular velocity
Joint i prismatic

Joint i revolute

39
 Geometric Jacobian – Computation
Column of geometric Jacobian

40
 Geometric Jacobian –
Representation in a Different Frame
 The Jacobian matrix depends on the frame in which the
end-effector velocity is expressed
 The above equations allow computation of the geometric
Jacobian with respect to the base frame
 For a different Frame t:

41
 RECAP: Geometric Jacobian –
Contribution of a Prismatic Joint
Note: joints beyond the i-th one are considered to be “frozen”,
so that the distal part of the robot is a single rigid body

42
 RECAP: Geometric Jacobian –
Contribution of a Revolute Joint
Note: joints beyond the i-th one are considered to be “frozen”,
so that the distal part of the robot is a single rigid body

43
 RECAP: Geometric Jacobian
It is possible to show that the i-th column of the Jacobian can be computed as

44
GJ-Example: 2 DOF manipulator

45
GJ-Example: 2 DOF manipulator

46
 GJ-Example: 2 DOF manipulator

maximum rank is 2  at most 2 components of the

linear/angular end-effector velocity can be independently assigned
47
GJ-Example: 3-link planar manipulator

48
GJ-Example: 3-link planar manipulator

49
GJ-Example: 3-link planar manipulator

50
GJ-Example: 3 DOF anthropomorphic manipulator

51
GJ-Example: 3 DOF anthropomorphic manipulator

52
GJ-Example: 3 DOF anthropomorphic manipulator

53
GJ-Example: 3 DOF anthropomorphic manipulator

54
GJ-Example: 3 DOF spherical manipulator

55
GJ-Example: 3 DOF spherical manipulator

56
GJ-Example: 3 DOF spherical manipulator

57
GJ-Example: 3 DOF spherical wrist

58
GJ-Example: PUMA 560

59
GJ-Example: PUMA 560

60
GJ-Example: Stanford manipulator

61
GJ-Example: Stanford manipulator

62