Error 10327: Can't determine definition of operator ""-""

I don't think you can do arithmetic within a bit selection. Probably something more structured like having a combinatorial signal that gets (47 - count). (Of course count is std_logic_vector and 47 is not, so you'll need to figure that out, probably writing 47 in binary "101111"). Then you'll probably need a for-loop that checks each bit of count and assigns and compares to the value, and if they match assigns RXD to that accx bit.

I've also heard recommended to use numeric_std libraries, which are more of a standard, instead of signed/unsigned/etc.(which are from Synopsys, although pretty much everyone understands them. I just think numeric_std does a better job of conversions operator overloading, although not positive).