Vector projection onto a straight line

Click For Summary
SUMMARY

The discussion focuses on determining the matrix for the spatial projection perpendicular to the line defined by the vector (1, 2, 3). The solution involves using the concept of linear projections, where the projection matrix is derived from the formula for projecting an arbitrary vector onto the line. The final projection matrix is calculated as (1/14) multiplied by the matrix [[1, 2, 3], [2, 4, 6], [3, 6, 9]]. This approach utilizes geometric intuition and the properties of dot products to arrive at the solution.

PREREQUISITES
  • Understanding of linear algebra concepts, specifically vector projections
  • Familiarity with matrix operations and properties
  • Knowledge of dot products and their geometric interpretations
  • Basic grasp of orthonormal vector spaces
NEXT STEPS
  • Study the Gram-Schmidt process for orthonormalization of vectors
  • Learn about the properties of projection matrices in linear algebra
  • Explore applications of vector projections in physics and engineering
  • Investigate the derivation and use of the least squares method in projections
USEFUL FOR

Students and professionals in mathematics, physics, and engineering, particularly those dealing with vector analysis and linear transformations.

Combinatus
Messages
40
Reaction score
1

Homework Statement



Determine the matrix for the spatial projection perpendicular to the straight line (x1, x2, x3) = t(1, 2, 3). The vector space is orthonormal.

Homework Equations


The Attempt at a Solution



After a trip to #math on freenode that resulted in discussions of Gram-Schmidt processes and bra-ket-like vector notations used in physics, I don't even know where to begin any longer. I suspect that this is supposedly a very rudimentary problem of geometric intuition, since I hadn't heard of projections until yesterday.

Anyway, if u=(a1,a2,a3) is an arbitrary vector and P is the projection (which is assumed to be linear), then P(u) = a1*P(e1) + a2*P(e2) + a3*P(e3), since e1, e2 and e3 are base vectors. Now, P(e1), P(e2) and P(e3) should be column vectors in the sought matrix.

Unfortunately, I don't know how to determine their coordinates.

In one attempt to determine them, I assigned s = (1,2,3) as a vector parallel to the line t(1,2,3). So, the dot product e1*s = |e1||s|cos(A) = 1*1 + 2*0 + 3*0 = 1, if e1=(1,0,0) and A is the angle between e1 and s. Then, A = arccos(1/sqrt(14)), since |s|=sqrt(14).

However, I'm getting nowhere with this approach. I can't really determine a normal of the line t(1,2,3) either. Bleh. Ideas are welcome.

Edit: Perhaps I can use the shortest distance between the tip of the vector u and the line...
 
Last edited:
Physics news on Phys.org
I think I solved it. Not sure if my approach was optimal, but oh well.

Basically, the sought projected vector should be parallel to s = (1,2,3), e.g. be written on the form L(1, 2, 3), where L is an unknown constant. Then, the vector perpendicular to the line between the tip of u and the line is then b = (L-a1, 2L-a2, 3L-a3). Using the dot product, we get b*s = L-a1 + 4L-2a2 + 9L-3a3 = 0 <=> L = (1/14)*(a1 + 2a2 + 3a3).

Inserting these values of L gives the coordinates of the sought vector as (1/14)*(a1 + 2a2 + 3a3)*(1,2,3). This corresponds to the projection matrix

<br /> \frac{1}{14}<br /> \begin{bmatrix}<br /> 1 &amp; 2 &amp; 3 \\<br /> 2 &amp; 4 &amp; 6 \\<br /> 3 &amp; 6 &amp; 9<br /> \end{bmatrix}<br />
 

Similar threads

  • · Replies 3 ·
Replies
3
Views
8K
  • · Replies 12 ·
Replies
12
Views
2K
  • · Replies 26 ·
Replies
26
Views
3K
  • · Replies 4 ·
Replies
4
Views
3K
Replies
5
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 11 ·
Replies
11
Views
20K
  • · Replies 2 ·
Replies
2
Views
2K