Simulink: 4 bytes to float

Hello!

I am sending with a microcontroller a float value (i.e. 145.2322308) as 4 bytes over Serial. I get the data in Simulink with " Serial Receiver" but I don't know how to combine the 4 bytes I am receiving in one variable of 32 bits and be able to scope the value 145.2322308.

With each serial package I send three values [float , float , long \CR]. I have tried different approaches but I can only scope the value of the bytes:

Thank you in advance for your help!

Answer 1

[Jonas]

Another way is to construct your output manually with 'Shift Arithmetic' blocks and perform bitshifts:

The input to the left you see is a vector of 4 bytes of type uint8.

Inside the MATLAB function is the following code, since no block exists to perform a typecast:

function single_value = typecast_int32_to_single(int32_value)
single_value = typecast(int32_value,'single');

[Walter Roberson]

If you are going to use a MATLAB Function block, then you can just typecast() the block of 4 bytes into single() directly; you might need to swapbytes() afterwards.

[Jonas]

@Walter Roberson Yes, also works! I tested and found that the swapbytes() was necessary as you mentioned.

So the MATLAB Function becomes:

function single_value = typecast_int32_to_single(uint8_buffer)
single_value = swapbytes(typecast(uint8_buffer,'single'));

I do find the generated code of my first suggestion a bit nicer:

compared to the integrated MATLAB Function:

So personally I am going to keep using my implementation with the bitshifts because I generate code from it.

[carlos munoz]

Thank you very much. It works!

Answer 2

[Jonas]

Use the Byte Unpack block and feed those 4 bytes into it after you have muxed them. Set Output port data type to {'single'} if you want to reconstruct a 32-bit float.