The error message is saying in essence that uint16 is different from fixdt(0, 16, 0) , which it is. Do you have uint16 data somewhere, perhaps feeding into the MATLAB Function block? If so, you should be able to fix this error by changing all the data types, both in- and outside the function block, to one or the other of the data types, as your needs require.
If you don't need fimath, Simulink will simulate faster if you use uint16. You can do this in the Function block with e.g. uint16(0) wherever yo need to specify the type, as opposed to fi(0, 0, 16, 0). The choice of data type will have no real effect on the generated HDL code.
