1: \begin{abstract}

2: Finite State Machine (FSM) model is widely used in the construction of binary convolutional codes.

3: If $\Z_2=\{0,1\}$ is the binary mod-2 addition group and $\Z_2^n$ is the $n$-times direct product of $\Z_2$,

4: then  a binary convolutional encoder, with rate

5: $\frac{k}{n}< 1$ and memory $m$, is a FSM with $\Z_2^k$ as inputs group, $\Z_2^n$ as outputs group 

6: and   $\Z_2^m$ as states group. The next state mapping $\nu:\Z_2^k\oplus \Z_2^m\rightarrow \Z_2^m$ is a surjective

7: group homomorphism. The encoding mapping $\omega:\Z_2^k\oplus \Z_2^m\rightarrow \Z_2^n$ is a homomorphism adequately

8: restricted by the trellis graph produced by $\nu$. The binary convolutional code is the

9:  family of bi-infinite sequences produced by the binary 

10: convolutional encoder. Thus, a convolutional code can be considered as a dynamical system 

11: and it is known that well behaved 

12: dynamical systems must be necessarily  controllable.

13: 

14: The generalization of binary convolutional encoders over arbitrary finite groups is made by  using the extension

15: of groups, instead of direct product. 

16: In this way, given finite groups $U,S$ and $Y$, a wide-sense homomorphic encoder (WSHE) is a FSM 

17: with $U$ as inputs group, $S$ as states group, and $Y$ as  outputs group.  By denoting $U\boxtimes S$ as the extension

18: of $U$ by $S$, the next state homomorphism  $\nu:U\boxtimes S\rightarrow S$ needs to be surjective 

19: and the encoding homomorphism  $\omega:U\boxtimes S\rightarrow Y$ has restrictions given by the trellis graph produced by

20: $\nu$. The code produced by a WSHE is known as group code.

21: In this work we will study the case when the extension $U\boxtimes S$ is abelian with $U$ being $\Z_p$, $p$ 

22: a positive prime number. We will show that this class of WSHEs  will produce controllable codes

23: only if the states group $S$ is isomorphic with $\Z_p^j$, for some positive integer $j$.

24: \end{abstract}

25: