One option:
x = (-20:0.05:20); %-20<x<20
y=5*(x.^2).*(sin(x));
yz = y.*circshift(y, [0 -1]);
yzi = find(yz < 0);
for k1 = 1:size(yzi,2)-1
xzeros(k1) = interp1(y(yzi(k1):yzi(k1)+1),x(yzi(k1):yzi(k1)+1),0);
end
figure(1)
plot(x, y, '-m')
hold on
plot(xzeros, zeros(size(xzeros)), 'pb')
hold off
grid
The routine finds the indices near the zero-crossings by multiplying the y-values with a one-index circularly-shifted version of itself and then finds the values where these products are negative. It then uses interp1 and the intervals defined by these indices to find the actual values of the zeros, producing the vector ‘xzeros’ that are the x-coordinates of the zeros. It then plots them.
