matlab signals

howerpower

United States619 Posts

August 09 2011 21:23 GMT

I'm taking a signals and systems course this summer and the professor sprang this project on us near the end of semester without ever introducing us to matlab. I've used the program before but never for a big project like this, although I think I'm learning it rather quickly.
I am however, getting stumped here.

I'm trying to code this filter in matlab, I know to define t as t=linspace(-pi/2,pi/2,17) but I guess what is confusing me the most is the part about defining the value of K.

Consider the ﬁlter {hn} which is deﬁned by the sampled values of the cosine wave
h(t) = K cos(t), t ∈ (−π/2, π/2),
at 17 points uniformly distributed in the interval (−π/2, π/2). Deﬁne value of K from the condition
Σh_n = 1

Count9

China10928 Posts

August 09 2011 21:30 GMT

I'm taking a signals and systems course this summer and the professor sprang this project on us near the end of semester without ever introducing us to matlab

Wait what? You guys really went through a whole semester of signals and systems without any matlab models? What school is this?

awu25

United States2003 Posts

August 09 2011 21:31 GMT

Isn't K the amplifying constant or something? I vaguely remember doing this in my Signals class. If you increased K, it would increase the response time or something, but at the same time cause a lot of overshoot. Don't really remember though, but I remember putting in random values in MATLAB for K. It sounds like you're trying to solve for a specific K value

josemb40

Peru611 Posts

August 09 2011 21:31 GMT

oh god, matlab, i hated that program. Can't you email your professor asking him for some kind of guide or advice about the project?

howerpower

United States619 Posts

August 09 2011 21:37 GMT

h=cos(t)
h=h/sum(h)

output is now
sum(h)=1

...profit?
I'm not sure if this is right but it does make sum(h)=1 lululul

actually I'm pretty positive it's right...guess I just needed to visit teamliquid to get my brain to work properly

Horuku

United States405 Posts

August 09 2011 21:44 GMT

What kind of filter are you talking about? At what frequency range do you need the amplification to be = 1?

forSeohyun

504 Posts

August 09 2011 21:51 GMT

Isn't it just Σh_n = ΣK cos(pi/(2*17)*n - pi/2) = 1 => K = 1/ ΣK cos(pi/(2*17)*n - pi/2)? n=0...17

This is because the gain of the filter is Σ|h(n)|^2 and you probably want that to be 1.

Or am I being stupid?

(I have read a couple of signals and systems/dsp/control courses)

TwistedHelix

United States50 Posts

August 09 2011 21:55 GMT

I'm taking a swing at this.
Converting the signal from cont. to discrete:

sum(h{n})= k sum[from n=-8 to 8] cos(n * pi /16)

The logic here is 1 period of the time domain is 2 * pi, and 1 period is 32 samples (16 in half a cycle, the 17 includes a point in the second half, which would mean h{n=-8}=0, and h{n=8}=0... thats how i understand it).
Sorry my sum doesnt start at 0, you can shift if you like.
Note our domain is the positive part of the cosine graph.
You should just be able to set the equation equal to 1, then evaluate.

1= k sum[from n=-8 to 8] cos(n * pi /16)

edit: changed wording

Primadog

United States4411 Posts

August 09 2011 23:45 GMT

A bit rusty, but I think K=2. The best way is to just test the results and see what you get.

obesechicken13

United States10467 Posts

August 10 2011 02:42 GMT

#10

My partner did everything, I only tested so I don't really know >.<

Please or register to reply.

matlab signals

Completed

Ongoing

Upcoming