Abs and angle of complex number with symbolic object

Hi,

If r=sym('a')*exp(j*sym('b')) is a symbolic object, and

1. if I write abs(r), I get abs(a).....I need get only "a" without abs

2. if I write angle(r), I get error

??? Undefined function or method 'atan2' for input arguments of type 'sym'.

Error in ==> angle at 14 p = atan2(imag(h), real(h));

Can you help me please? How can I solve this?

Answer 1

[Walter Roberson]

There is no "complex argument" function for symbolic variables exposed at the MATLAB level. You can, though, define

arg = @(V) feval(symengine, 'arg', V);

http://www.mathworks.com/help/symbolic/mupad_ref/arg.html

[john]

Sorry, I realy don't understand, what makes code arg = @(V) feval(symengine, 'arg', V);....what means V and how can I use that code

[Walter Roberson]

http://www.mathworks.com/help/matlab/matlab_prog/anonymous-functions.html

After the definition above, you would use

arg(r)

[john]

Did you mean this code

r=sym('a')*exp(j*sym('b'));

arg = @(r) feval(symengine, 'arg', r);

number=arg(r);

????

But I got again error

??? Undefined function or method 'atan2' for input arguments of type 'sym'.

Error in ==> angle at 14 p = atan2(imag(h), real(h));

Btw, what is better to use, atan or atan2 in form "P = 2 * atan( (sqrt(X.^2 + Y.^2) - X) ./ Y )";

Thank you

Answer 2

[Brian B]

Well, for part 1 you have to assume a is a non-negative real:

 syms a
 assume(a>0)

Then

 >> abs(a)
 ans =
 a

I have not tried the second part, but you can imagine dividing by the magnitude and taking a log....

[Brian B]

You may also have to

 assume(b,'real')

[john]

For code

 syms a
 assume(a>0)

I got error

??? Error using ==> sym.sym>notimplemented at 2653
Function 'gt' is not implemented for MuPAD symbolic objects.

Error in ==> sym.sym>sym.gt at 801
            notimplemented('gt');

[Brian B]

What version of the Symbolic Math Toolbox do you have?

 ver('symbolic')

[john]

Hi,

Symbolic Math Toolbox Version 5.6 (R2011a)

[Brian B]

Try

 a = sym('a','positive')
 b = sym('b','real')

That works for me in R2011b, including computing the magnitude of r. I don't see a good way to get the argument, though.

[john]

Tanks, but It doesn't help for angle(r)