Calculate $d$ from $[n,k,d]$-Code given a parity check matrix

293 Views Asked by At

Let $C$ be the $[n,k,d]$-Code with the parity check matrix B in $\mathbb{F}_2$ with parity check matrix (sry, don't want to type that thing in Latex).

I already found out that $n=15, k=11$. For $d$ my idea was to calculate $C$ first, by solving $Bx=0$ and then calculate $d(C)$. My result is $d(c)=3=d$, is this true, or am I doing something wrong here?

Extra material

(Here you can see the solved system with matrixcalc)

1

There are 1 best solutions below

2
On

Well, correct me if I'm wrong, but the parity check matrix contains each nonzero binary vector of length 4 as a column vector (15 vectors in total). This is a binary Hamming code with easy decoding.

In view of minimum distance, for each codeword $c\ne 0$, $Bc^t=0$ (transposition). But $Bc^t$ is a linear combination of the columns of $B$ (the columns at which $c$ has the nonzero entries). Such a nontrivial linear combination can only be zero if three or more columns are combined due to the structure of the columns of $B$. So the Hamming weight of $c$ is at least 3. Moreover, there exist three columns of $B$ that add up to zero and so the minimum Hamming weight of the code is indeed 3, i.e., the minimum distance is $d=3$,

In view of decoding, if $c$ is the vector sent, $y$ is the vector received, and $e=y-c$ is the error vector, then $Be^t = By^t-Bc^t = By^t$. If there is one error, then $By^t = Be^t\ne 0$ is a column of $B$. If its the $i$th column vector, then the error can be corrected by flipping the $i$th component in $y$.