symbolic equation Empty sym: 0-by-1

I'm newbie to matlab and when i tried to solve this equations it gives me an empty struct field

syms osx osy osz oex oey oez lf ls a k3 xe ye ze ya2 Rd Td ls

Td=[22;35;63];

lf=0.25;

Rd=[10 20 30 ;40 50 60 ;70 80 90];

ls=0.3

%%given rotation matrix and postion vector

os= Td-Rd*[0;0;1].*ls;

%%%%

a=atand(ye/ze);

k3=atand(xe/(cos(a)*ze-sin(a)*(ye-ya2)));

oez=os(3)-lf*(cos(a)*cos(k3));

oey=os(2)-lf*(-sin(a)*cos(k3));

oex=os(1)-lf*(sin(k3));

%%%

sol=solve([oex,oey,oez],[xe,ye,ze])

ny help appreciated..Thanks

[KALYAN ACHARJYA]

Undefined function 'atand' for input arguments of type 'sym'.
Error in test_code1 (line 12)
a=atand(ye/ze);

???

[John D'Errico]

@Kalyan - you may have an older MATLAB release, that does not support atand operating on symbolic arguments.

My R2019a release has no such problem.

atand(xe)
ans =
(180*atan(xe))/pi

[KALYAN ACHARJYA]

Thanks @John for the useful information.

Answer 1

[John D'Errico]

When solve returns an empty result, that means it can find no analytical solution.

And since you have symbolic parameters in the equations, no numerical solution can exist. That is, at least ya2 has no value set, so it is a symbolic parameter. There might be others, but I did not look carefully for more than one parameter.

Sorry, but it is pretty easy to pose a problem for which no solution exsts. In fact, you just did it!

If you have a value for ya2 (and any other parameters in there that I missed) then you can use solve. The solution will probably not be unique then. And even then, it might only exist as a numerical solution.

Just because you want something to exist, does not make it so. Unless of course, your wand skills are pretty good. Mine simply are not sufficient.

[abdelrahmn ahmed]

I entered all values and one of the equation i want to solve is :

oex =

13 - sin((180*atan(xe/(ze*cos((180*atan(ye/ze))/pi) - sin((180*atan(ye/ze))/pi)*(ye - 1/2))))/pi)/4

I have three equations (oex,oey,oez)with three variables(xe,ye,ze).

why it won't be solved?