||
151-0851-00 Vlecture: CAB G11 Tuesday 10:15 – 12:00, every weekexercise: HG G1 Wednesday 8:15 – 10:00, according to schedule (about every 2nd week)office hour: LEE H303 Friday 12.15 – 13.00Marco Hutter, Roland Siegwart, and Thomas Stastny
20.09.2016Robot Dynamics - Kinematics 1 1
Lecture «Robot Dynamics»: Kinematics 1
||
Builds upon notation of other dynamics classes at ETH and IEEE standards
Vector: (often also ) Vector from point B to P: Reference coordinate system (calligraphic)
:= orthonormal basis of R3
Numerical representation of a vector: Addition of vectors: Use the same reference frame:
20.09.2016Robot Dynamics - Kinematics 1 2
Recapitulation: Vectors, Position, and Vector Calculus
r r
BPr{ }A
BPrA
AB
P
AP AB BP r r rABAP BP rr rA A A
BPrAPr
ABr
{ }A
||
Cartesian coordinates
Position vector
Cylindrical coordinates
Position vector
Spherical coordinates
Position vector 20.09.2016Robot Dynamics - Kinematics 1 3
Parameterization of Vectors
||
Only for Cartesian coordinates it holds that
NEVER do this for other representations (requires special algebra!!) => we will encounter similar problems for rotations
20.09.2016Robot Dynamics - Kinematics 1 4
Parameterization of VectorsExample
1 1 01 0 11 0 1
AP AB BP
r r rA A A
10 :0
Pc
Pz
Ps
χχχ
01 :1
Pc
Pz
Ps
χχχ
11 :1
Pc
Pz
Ps
χχχ
AP AB BP χ χ χ
1,0,0
1,0,0
1,0,0
T
T
T
0,1,1
1, ,12
2, ,2 4
T
T
T
1,1,1
2, ,14
13, ,arctan4 2
T
T
T
||
The velocity of point P relative to point B, expressed in frame A is:
Question: What is the relationship between the velocityand the time derivative of the representation
20.09.2016Robot Dynamics - Kinematics 1 5
Differentiation of Representation Linear Velocity
BPrA
rχ
r r χrr χχ
1
P
P
r E χ χ
χ E χ r
||
Cartesian coordinates:
Cylindrical coordinates:
20.09.2016Robot Dynamics - Kinematics 1 6
Differentiation of Representation Linear Velocity
||
Position of P with respect to A expressed in A:
Position of P with respect to A expressed in B:
Write the unit vectors of B expressed in A as:
=>
20.09.2016Robot Dynamics - Kinematics 1 7
Rotations
||
The rotation matrix transforms vectors expressed in B to A:
The matrix is orthogonal:
Belongs to special orthonormal group SO(3) (and not R3) This causes difficulties and requires special algebra
Consecutive rotations:
20.09.2016Robot Dynamics - Kinematics 1 8
Rotation Matrix
||
Passive rotation = mapping of the same vector from frame {B} to {A}
Active rotation = rotating a vector in the same frame
20.09.2016Robot Dynamics - Kinematics 1 9
Passive and Active Rotation
R
||
Find the elementary rotation matrixs.t
20.09.2016Robot Dynamics - Kinematics 1 10
Elementary Rotation
|| 20.09.2016Robot Dynamics - Kinematics 1 11
Homogeneous TransformationCombined Translation and Rotation
Homogeneous transformation = translation and rotation
Inverse
||
This allows to transform an arbitrary vector between different reference frames(classical example: mapping of features in camera frame to world frame)
20.09.2016Robot Dynamics - Kinematics 1 12
Homogeneous TransformationsConsecutive Transformation
{A}{B}
AB
P
{C}
CAP BP
BP CP
r rr r
TT T T
T
A AB B
AC AB BCB BC C
||
Find the position vector Find the transformation matrix
Find the vector
20.09.2016Robot Dynamics - Kinematics 1 13
Homogeneous TransformationSimple Example
1 0 0 00 0 1 30 1 0 10 0 0 1
TAB
1 1
1 0 0 0 0 00 0 1 3 1 20 1 0 1 1 20 0 0 1 1 1
AP BPr r
T
A BAB
||
Angular velocity describes the relative rotational velocity of B wrt. Aexpressed in frame A
The relative velocity of A wrt. B is: Given the rotation matrix between two frames, the angular velocity is
Transformation of angular velocity:
Addition of relative velocities: 20.09.2016Robot Dynamics - Kinematics 1 14
Angular Velocity
ω ωAB BA
||
Given the rotation matrix
=>
20.09.2016Robot Dynamics - Kinematics 1 15
Angular VelocitySimple Example
||
Rotation matrix: 3x3 = 9 parameters Orthogonality = 6 constraints
Euler Angles 3 parameters singularity problem
Angle Axis 4 parameters (angle and axis) unitary constraint
Quaternions 4 parameters no singularity constraints
20.09.2016Robot Dynamics - Kinematics 1 16
Outlook (next week)Rotation Parameterization