problems with "if statement"

Hi everyone. I am beginning in Matlab and I have a problem.

My code follows:

for i=1:Nped
    for j=1:length(t)
        dist(i,j)=vp(i)*t(j);
        for k=1:NH       
            if (dist(i,j)<=Lenght) 
                Findiv(i,:)=G(i)+alfadin(1)*G(i)*sin(2*pi*i*fp(i)*(t-Tentr(i))-phase(1));
            else
                Findiv(i,:)=0
            end
        end
    end

For some reason, the loop :

if (dist(i,j)<=Lenght)

.... is not been well executed. I ploted the graphic of dist(1,j) vs t (time) and I verified that for a distance dist(1, j ) > lenght (where lenght= 20 m), the non-zero expression give by

 Findiv(i,:)=G(i)+alfadin(1)*G(i)*sin(2*pi*i*fp(i)*(t-Tentr(i))-phase(1))

is not calculated (as showed in the second graph).

Does anyone could help me to solve this problem?

Thanks a lot =)

[Image Analyst]

What is Lenght, lenght, and length? Those are three different spellings and capitalizations and MATLAB is sensitive to mispellings and is case sensitive. Also, where is your plotting code?

[Torsten]

Where do the loop variables "k" and "j" come into play within the k-loop ?

As written, the k-loop is superfluous.

[Igor Braz Gonzaga]

I'm sorry, I copied the test code. The correct follows:

 for i=1:Nped
        Findiv=G(i)+alfadin(1)*G(i)*sin(2*pi*i*fp(i)*(t-Tentr(i))-phase(1));
        for j=1:length(t)
            dist(i,j)=vp(i)*t(j);
        for k=1:NH       
            if (dist(i,j)<=20) %&(t(j)>=Tentr(i))
                Findiv(i,:)=G(i)+alfadin(k)*G(i)*sin(2*pi*i*fp(i)*(t-Tentr(i))-phase(k));
            else
                Findiv(i,:)=0;
            end
        end
            %Ft(:)=Ft(:)+Findiv(i,:)
    end

[Jan]

In this version "phase(k)" appears. In the version you post in the comments below, it is "pahse(1)". This optimizes the cofusion fro the readers. We see 3 different codes now and all of them do, what they are doing. You are expecting something else, but we cannot guess why.

Use the debugger to examine whats going on. Set a breakpoint in the first line, then step through the code line by line. Of course, Matlab does exactly what it is instructed to do. The IF condition does work properly, so the only problem is that you expect something else.

The purpose if the "for k=1:NH" loop is still not clear and inside it Findiv is overwritten repeatedly. This cannot be useful.

Answer 1

[Jan]

The if command is not a loop.

What is the purpose if the "for k" loop? Its body does not depend on k, so you can omit this loop. Instead of setting the value of Finfiv explicitly to 0, you can pre-allocate also:

Findiv = zeros(Nped, 1);
for i = 1:Nped
    tmp = G(i) + alfadin(1) * G(i) * ...
          sin(2 * pi * i * fp(i) * (t - Tentr(i)) - phase(1));
    for j = 1:length(t)
        dist(i,j) = vp(i) * t(j);     
        if dist(i, j) <= Lenght
            Findiv(i) = tmp;
            break;
        end
    end
end

I've moved the calculation of G(i) + alfadin(1) *... out of the inner loop, because it does not depend on j and calculating it repeatedly is a waste of time. Therefore I've inserted a break also to leave the inner loop soon.

Maybe this is the problem: Findiv(i) is overwritten in each iteration of the loop over j, so it depends on the last value vp(i)*t(end) only.

[Torsten]

tmp is a vector because of the t-array in it.

[Igor Braz Gonzaga]

Hi Jan, I'm sorry. I coppied I test, the correct code follows:

for i=1:Nped
    Findiv=G(i)+alfadin(1)*G(i)*sin(2*pi*i*fp(i)*(t-Tentr(i))-phase(1));
    for j=1:length(t)
        dist(i,j)=vp(i)*t(j);
        for k=1:NH       
            if (dist(i,j)<=20) 
                Findiv(i,:)=G(i)+alfadin(k)*G(i)*sin(2*pi*i*fp(i)*(t-Tentr(i))-phase(k));
            else
                Findiv(i,:)=0;
            end
        end
    end

[Image Analyst]

After you read this link

TUTORIAL: How to ask a question (on Answers) and get a fast answer

Tell us the missing variables (like Nped, alfadin, etc.) so we can actually run your code. And, again, show us the code you used to plot those two figures you posted.

[Igor Braz Gonzaga]

Nped is an integer = 6 (number of pedestrians);

alfadin is a vector with two components: alfadin(1)=0.4 and alfadin(2)=0.1;

Lenght=20m is the controller variable that control Findv, if L<= 20m, than Findiv is calculated, else, Findiv=0;

dist( Nped, Lenght(t)) = is the distance. Calculated by multiplying the velocity of each person "v(i)" by the time " t ";

To plot the graphs I used the lines on the end of the code

for i=1:Nped
    for j=1:length(t)
        dist(i,j)=vp(i)*t(j);
        for k=1:NH       
            if (dist(i,j)<=Lenght) 
                Findiv(i,:)=G(i)+alfadin(1)*G(i)*sin(2*pi*i*fp(i)*(t-Tentr(i))-phase(1));
            else
                Findiv(i,:)=0;
            end
        end
    end     
    plot(t,Findiv(1,:));
    xlabel('Time');
    ylabel('Findiv');

[Image Analyst]

Nped = 6;
alfadin(1)=0.4 ;
alfadin(2)=0.1;
Lenght=20;
for i=1:Nped
    for j=1:length(t)
        dist(i,j)=vp(i)*t(j);
        for k=1:NH       
            if (dist(i,j)<=Lenght) 
                Findiv(i,:)=G(i)+alfadin(1)*G(i)*sin(2*pi*i*fp(i)*(t-Tentr(i))-phase(1));
            else
                Findiv(i,:)=0;
            end
        end
    end     
    plot(t,Findiv(1,:));
    xlabel('Time');
    ylabel('Findiv');
end

Unrecognized function or variable 't'.

So is that what you get?

[Igor Braz Gonzaga]

Yes. But something is wrong. Take a look what happens when I plot Findiv(1, : ):

I've plotted the graph of dist( 1, : ) and it works well (it means that there is a limite, in time, between values smaller and higher than the defined parameter Lenght, as indicated in the law of the IF condition).

So, I dont understand why Findv (1 , :) is zero for all times. It is like the condition "IF" just works for the ELSE block.

[Torsten]

Make that MATLAB recognizes t - otherwise the graphs will make no sense.

Still, the k-loop is superfluous - k is never referenced.

[Image Analyst]

You say that's what yhou got but that's not true. If you did get that, then you'd have gotten the error about t not being defined and you never would have gotten your plot. I'm really about to quit here. Either post the full code or I'm outta here. Perhaps Torsten will still be kind enough to help you.

Answer 2

[Voss]

A couple of people have pointed out that the "for k=1:NH" loop is not necessary, but since it is still part of the code, can you check the value of NH?

If NH is less than 1 or NH is the empty array, then the "for k=1:NH" loop will execute zero times, leading to the if statement not being checked at all and Findiv not being calculated (but dist will be calculated because that is done outside the "for k=1:NH" loop).

This possibility is consistent with your observations of the values of dist and Findiv, if Findiv is pre-allocated with zeros.