Engineering Mathematics - 4
Project on
Orthogonal Transformations
Submitted By
1. [Your Name 1]
2. [Your Name 2]
3. [Your Name 3]
4. [Your Name 4]
Under the Guidance of
Prof. [Faculty Name]
Department of Computer Science and Engineering
(Artificial Intelligence and Machine Learning)
SARASWATI COLLEGE OF ENGINEERING
INTRODUCTION
An Orthogonal Transformation is a linear mapping of vectors that maintains
the angles and lengths of vectors. In simpler terms, it preserves the
Euclidean structure of vector spaces. These transformations include
rotations, reflections, or combinations of both, without altering the shape or
size of geometrical figures.
In mathematics, such transformations are represented using orthogonal
matrices, where the transpose of the matrix is equal to its inverse: AᵀA = I,
where A is the transformation matrix. This concept is fundamental in areas
such as computer graphics, robotics, and signal processing.
PROPERTIES OF ORTHOGONAL
TRANSFORMATIONS
Preservation of Length: Any vector v transformed using an orthogonal
matrix A satisfies ||Av|| = ||v||.
Preservation of Angle: Inner products are preserved, i.e., if u and v
are vectors, then (Au)·(Av) = u·v.
Orthogonality of Columns: Columns (and rows) of orthogonal
matrices are mutually orthogonal and normalized.
Determinant: The determinant of an orthogonal matrix is always ±1.
Inverse = Transpose: Orthogonal matrices are easy to invert due to
this property.
Orthogonal transformations are distance-preserving and therefore
classified as isometries.
APPLICATIONS
Engineering & Robotics
Orthogonal transformations are widely used in the field of robotics
to perform coordinate transformations and simulate the movement
of robotic arms. These transformations help model rotational
motion without distortion, making them essential in kinematics and
dynamics analysis.
Computer Graphics
In 2D and 3D graphics, these transformations are used for rotating,
reflecting, and scaling models without losing proportions. Real-
time rendering of scenes and objects depends heavily on such
calculations.
Machine Learning & Data Science
In dimensionality reduction techniques like Principal Component
Analysis (PCA), orthogonal transformations are used to rotate the
feature space while maintaining the original data structure. This
helps improve model efficiency and interpretability.
Signal and Image Processing
Transformations such as the Fourier Transform and Discrete
Cosine Transform (DCT) are special types of orthogonal
transformations. They help in compressing signals and images
while retaining important information.
CONCLUSION
Orthogonal transformations play a critical role in mathematics and
engineering. They not only simplify calculations but also preserve the
fundamental structure of data. Their unique properties make them suitable
for a wide range of applications including robotics, computer vision,
quantum mechanics, and machine learning.
Understanding orthogonal transformations enhances our ability to model and
solve real-world problems efficiently. Their role as isometries makes them a
cornerstone in applied linear algebra and numerical computing.
NUMERICALS
Q1: Verify if a Given 4×4 Matrix is Orthogonal
A=[ 1 0 0 0 ]
[ 0 0 -1 0 ]
[ 0 1 0 0 ]
[ 0 0 0 1 ]
Soln:-
Step 1: Compute A^T (Transpose of A)
A^T = [ 1 0 0 0 ]
[ 0 0 1 0 ]
[ 0 -1 0 0 ]
[ 0 0 0 1 ]
Step 2: Compute A^T * A
Multiplying A^T and A:
[ 1 0 0 0 ] [ 1 0 0 0 ]
[ 0 0 1 0 ] x[ 0 0 -1 0 ]
[ 0 -1 0 0 ] [ 0 1 0 0 ]
[ 0 0 0 1 ] [ 0 0 0 1 ]
After performing the multiplication, we get the Identity matrix I:
I =[ 1 0 0 0 ]
[ 0 1 0 0 ]
[ 0 0 1 0 ]
[ 0 0 0 1 ]
Since A^T * A = I, the given matrix A is orthogonal.
Q2: Rotation of a Point Around the X-Axis
Rotate the point P(1,2,3) by 90° around the X-axis using the transformation matrix:
R= [ 1 0 0 ]
[ 0 cosθ -sinθ ]
[ 0 sinθ cosθ ]
Soln:-
Step 1: Substituting θ = 90°
Since cos(90°) = 0 and sin(90°) = 1, the rotation matrix becomes:
R= [ 1 0 0]
[ 0 0 -1 ]
[ 0 1 0]
Step 2: Compute the New Coordinates
Multiplying R_x with point P(1,2,3):
[ 1 0 0 ] [ 1 ] = [ (1×1) + (0×2) + (0×3) ] = [ 1 ]
[ 0 0 -1 ] x [ 2 ] = [ (0×1) + (0×2) + (-1×3) ] = [ -3 ]
[ 0 1 0 ] [ 3 ] = [ (0×1) + (1×2) + (0×3) ] = [ 2 ]
Thus, the new coordinates after rotation are (1, -3, 2).
Q3: Reflection across an Arbitrary Plane in 3D
Find the image of the point P(4, 2, 3) reflected across the xy – plane.
Soln:-
The reflection matrix for the XY- plane is:
R=[ 1 0 0 ]
[ 0 1 0 ]
[ 0 0 -1 ]
Applying the transformation:
P` = R x P = [ 1 0 0 ] [ 4 ]
[ 0 1 0 ] x [ 2 ]
[ 0 0 -1 ] [ 3 ]
=[ 4 ]
[ 2 ]
[ -3 ]
New coordinates: (4, 2, -3).
Q4: Projection of a Vector
Find the orthogonal projection of v = (3, 4) onto u = (2, 1).
Soln:-
The projection formula is:
v .u
proj v = u
u .u
Step 1: Compute dot products
v . u= (3 x 2 ) + ( 4 x 1 )=6+ 4=10
u . u=( 2 x 2 ) + ( 1 x 1 ) =4 +1=5
Step 2: Compute the projection
10
proj v = ( 2 , 1 )=2 ( 2, 1 )=(4 , 2)
5
Thus, the projection is (4, 2).