I'm calculating the square of a polynomial inside a loop. It seems to me that, calculating it using conv() is resulting in large errors, compared to converting the polynomial to symbolic and do the calculation there.
Just out of curiosity, I tried to sym2poly the symbolic results back to polynomial, and the error appears again.
However, I can't using symbolic calculation in the loop, because it's too slow.
Any input on why this is happening and how to solve it properly? Thanks in advance.
D4_poly= 1.0e+06 * [ -0.000001003321333
D4_fn = matlabFunction(poly2sym(D4_poly));
subplot(2,2,1);fplot(D4_fn,[19.999825010467916,23.158281544048581])
SqD4_poly = conv(D4_poly',D4_poly');
SqD4_fn = matlabFunction(poly2sym(SqD4_poly));
subplot(2,2,2);fplot(SqD4_fn,[19.999825010467916,23.158281544048581])
temp = poly2sym(D4_poly);
SqD4_s_fn = matlabFunction(temp ^2);
subplot(2,2,3);fplot(SqD4_s_fn,[19.999825010467916,23.158281544048581])
SqD4_poly_2 = sym2poly(temp ^2);
SqD4_3_fn = matlabFunction(poly2sym(SqD4_poly_2));
subplot(2,2,4);fplot(SqD4_3_fn,[19.999825010467916,23.158281544048581])
