on hw#5, I am not clear as to what is being asked for in problem 1a.
in hw#1 the hammer was at (6,0) and rotated (in frame B) -pi. in hw#3 the hammer was at (1,0) and rotated (in frame B) -pi. nowhere was the hammer rotated -pi/2.
I'll answer your question this evening.
Ok thanks
Dr. Vela is defining the relationship between the base of the hammer and the hammer head in the problem description. This hasn't really been done before. He is only saying that the transformation from the base to the head has a d of [1;0] and a rotation by -pi/2. Does that help?
Am I to reference any prior homework to solve 1a?
Is the problem asking me to rotate and translate the world frame, O, to the frame A (from earlier homework #3) and then to translate and rotate that frame to the location described in the homework and call that frame A'?
In homework 3, you are given the transformations from O to A and B, which are each the base of the hammer. Now in this homework you are given the transformation from the base of the hammer to the head. Question 1A is asking you to compute the overall transformation from O to the hammer head in A and from O to the hammer head in B.
1a has nothing to do with 'adjoint'
Nope not really
thanks I will try to continue on the rest this evening
Sounds good
I took gOA, I calculated in Hw3, and multiplied it with the hammer head point and rotation gOA*gHAMMER gHAMMER = ({1 0 }, R) and got a different answer than gHAMMER*gOA this tells me I am doing something incorrect in Hw5 1a
I am thinking of gHAMMER as a frame
never mind I got it
Yeah gOA*gHammer should not equal gHammer*gOA
in 2a is the response qpOA = ({0 0}, R(theta)) ?
or gAB = ({0 0}, R(theta))
gAB = ({x y}, R(theta)) where x and y are constants for all values of theta
none of that worked, how should I be approach 2a?
Ok so in 2A, you want to describe the following statement: A transformation of a certain point returns that same point.
the pole is the point in the plane whose coordinates do not change when the point is rigidly transformed by g
all I can think of is g = ({0 0}, and the rotation is any angle)
Hmmm ok start by writing out the equation of a transformation of a point by g.
Conceptually, we're not trying to figure out g...we're trying to figure out qp for a given g. So you want to write an equation taht relates g and qp, not a specific g
qp = gAB*qp
Right...now the only other thing we have to say is that the coordinates of the point do not change after the transformation...
Well actually I suppose you've already done that in your equation...so that's basically it
2b I have qp = d +R*qp but what does it mean by what frame was it computed in?
If you write superscripts and subscripts on the variables, the frame it's calculated in is the superscript over qp. However, problem 2 asks to solve for qp, meaning you can not use qp as part of it's definition. In other words, you have to do some algebra to isolate qp
qp = gAB*qBB the B's cancel leaving qp = qAB and the frame is A
What you've essentially written is qAp = gAB*qBp and that qAp = qBp...we already knew that. What you want to do is define qAp=qBp in terms of R and d of g. As the problem states, you should first write out your previous equations in terms of d and R.
You've done that here: qp = d +R*qp. Now just express qp only in terms of d and R...not in terms of itself.
if I do this qp = gAB*qp = d + R*qp qp - R*qp = d qp = d/(1 - R) now qp is in terms of d and R
Close but remember that you're dealing with matrices. For one thing, you can't divide matrices (you can only multiply by the inverse). And the other thing is that the side you multiply on is important so that inv(A)*A*B = B but A*B*inv(A) does not equal B.
how do you subtract a matrix from 1 as in (1 - R)?
The 1 in this case would be the identity matrix, with 1s on the diagonal and 0s everywhere else.
qp = d*inv(1 - R)
Almost there but when you factored out qp on the left side, I think you factored it out on the wrong side.
qp = -d*inv(R - 1)
Nope... what did you get when you factored out qp on the left side of the equation?
qp*(1-R)
qp = inv(1-R)*d
qp = gAB*qp = d + R*qp qp - R*qp = d (1 - R)*qp = d inv(1 - R)*(1 - R)*qp = inv(1 - R)*d qp = inv(1 - R)*d
yep that looks good to me
where 1 = I
what frame does it mean? in what frame ? since qp is the same in all frames, does a frame apply?
It's not the same in all frames, it's only the same in the two frames which the g relates. I think Dr. Vela is trying to make you see taht the point you found is in the frame A or frame B, but not in the observer frame.
for 3a do I multiply the vector by the rotation matrix to get the vector in relation to the frame?
Yes that's correct.
Join our real-time social learning platform and learn together with your friends!