Need to do a 8-Bit Fletcher Algorithm checksum.

I have to sent serial command to a gps receiver but it is necessary create 2 checksum bytes after all commands.

The receiver protocol is the 8-bit Fletcher.

The number of bytes in the message can be variable. The algorithme in an other language is:

%%%%%%%%%

CK_A = 0, CK_B = 0
For(I=0;I<N;I++)
{
CK_A = CK_A + Buffer[I]
CK_B = CK_B + CK_A
}

%%%%%%%%%%

I'would not prefare use the famous 'dec2bin' if possible

thank you

Answer 1

[Rushil]

Hi Patrice

From the provided implementation, I assume that the buffer is a sequence of 8-bit integers for which we need to compute the checksums. We can use MATLAB to easily accomplish this task using “mod” function (to retain the 8-bit nature of the integers). Below you will find implementation of the same:

function [CK_A, CK_B] = fletcher(Buffer)
    CK_A = uint8(0);
    CK_B = uint8(0);
    N = length(Buffer);
    for I = 1:N
        % use mod with 2^8 (which is 256)
        CK_A = mod(CK_A + uint8(Buffer(I)), 256);
        CK_B = mod(CK_B + CK_A, 256);
    end
end

% an example of using it
Buffer = [1, 2, 3, 4, 5, 6];
[CK_A, CK_B] = fletcher(Buffer)

To help with understanding the function, you may find the documentation of “mod” below:

https://mathworks.com/help/matlab/ref/double.mod.html

Hope it helps you out

larush

[Walter Roberson]

Not quite. You are adding two uint8 values. When the result of adding two uint8 values would overflow uint8 the result saturates to become uint8(255) . There is no point using mod(256) on the addition of two uint8 values because the result always saturates at most 255. Which is not what you want happening.

So instead you need to use uint16 for the calculations.

function [CK_A, CK_B] = fletcher(Buffer)
    CK_A = uint16(0);
    CK_B = uint16(0);
    N = length(Buffer);
    for I = 1:N
        % use mod with 2^8 (which is 256)
        CK_A = mod(CK_A + uint16(Buffer(I)), 256);
        CK_B = mod(CK_B + CK_A, 256);
    end
    CK_A = uint8(CK_A);
    CK_B = uint8(CK_B);
end

% an example of using it
Buffer = [1, 2, 3, 4, 5, 6];
[CK_A, CK_B] = fletcher(Buffer)

This code does assume that the input Buffer has no elements greater than 255.

[Rushil]

Hi Walter

Thanks for the response. I may have overlooked this fact, since I am used to using macros in C++.

#define int long long

Usually most operations (multiplication too) performed modulo some 32-bit number work, since the 64-bit operations allow the operations on large integers, which would otherwise overflow. I guess this is a good opportunity for me to learn more about this in MATLAB.

Your comment helps not only 1, but 2 people, thank you for the clarification.

larush

[patrice boisset]

hi thanks for all