Convolution code generator and digital signal processor...

Details Convolution code generator and digital signal processor... Convolution code generator and digital signal processor...

1999-06-01

2002-08-27

Decady, Albert (Department: 2133)

Error detection/correction and fault detection/recovery

Pulse or data error handling

Digital data error correction

Reexamination Certificate

active

06442729

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to convolution code generation, and more particularly to a convolution coder of a digital signal processor which consists of a hardware in the digital signal processor and provides a plurality of convolution codes by a single instruction.
2. Description of the Background Art
Data must be transmitted to a receiving part without any error in transmission and reception of digital data. However, since errors can be made in general during the data transmission and reception, many ways have been proposed and actually used to prevent the errors. These can be referred to as an error correction code technique, which includes a block coding and a convolution coding. The present invention relates to the convolution coding technique, in particular, which has been recently used for cellular phones.
To process digital signals, a digital signal processor (DSP) is exclusively used. The DSP has a function of convolutely coding data and supplying the coded data to a digital data receiving part, to transmit the data without error employing the digital processor which provides numerous functions related to the digital signal processing. Now, the convolution coding performed in a conventional DSP will be described.
Generally, for bit-operation in the DSP there are provided instructions such as arithmetic shift or logical shift, AND, OR, EX-OR (exclusive OR), etc., and also there is provided an instruction for fetching a bit in a single word. Therefore, a convolution coding which is a specific signal processing with respect to an input signal is performed using the above-mentioned instructions for the bit operation supplied by the DSP.
Now, the convolution coding will be described with reference to
FIG. 1
which is a conceptional diagram. A convolution code with respect to inputted data x(n) is obtained as shown therein. Since an exclusive unit for such a convolution coding does not physically exist in the DSP, the convolution coding is processed using a plurality of instructions.
As shown therein, the convolution coding process, marked a referential numeral
1
, produces a convolution code with respect to the inputted data x(n) using a shift register
2
and exclusive logic elements
3
,
4
. In
FIG. 1
when expressing convolution codes n, k, m, these are assumed to be 2, 1, 5, respectively. Here, n is a code word length, that is, the number of bits which are encoded and outputted by the convolution coding each time step, which is 2 in accordance with the above example, k is a message block length which indicates the number of data inputted by each time step, wherein 1 bit is inputted each time step in the above example, and m is a memory block and indicates the number of shift registers in which the inputted data are stored, as shown in
FIG. 1
, which is required to use a register having 5 cells. x(n) indicates data inputted from a time step n.
Since the input data x(n) is shifted to the left to the right side of the shift register
2
each time step, a first cell D
1
, a fourth cell D
4
and a fifth cell D
5
of first to fifth cells D
1
-D
5
constituting the shift register
2
are exclusively logic-operated by the first exclusive logic element
3
, thereby generating a first output G
0
(n). In addition, a second output G
1
(n) is generated with respect to the first cell D
1
, the second cell D
2
, the fourth cell D
4
and the fifth cell D
5
by the second exclusive logic element
4
. Thus, the above processes can be expressed as follows.
G
0
(1)=
x
(
n
)+
x
(
n−
3)+
x
(
n−
4)
G
1
(1)=
x
(
n
)+
x
(
n−
1)+
x
(
n−
3)+
x
(
n−
4)  (1)
In the above example of the (2, 1, 5) convolution code process, when a single bit is inputted to the shift register
2
, a couple of outputs are generated by the exclusive logic operation with respect to a bit value in cells which are selected in the time step or a modular-2 addition. That is, whenever a single bit inputted, 2 bits are outputted.
However, since the instructions for the bit operation provided by the DSP must be used in the convolution coding to the input data stream, a sequential instruction performance is required. In other words, the instructions are used in the digital signal processor, the data stream given by sequential instructions becomes convolutely coded.
First, the bit stream constituting a data stream to be inputted is sequentially stored from a highest bit to a lowest bit in a specific part of a memory word by word. Accordingly, the memory includes, for example, a k number of words. Then, when the number of bits included in a single word is N, considering a general size, as shown in
FIG. 1
, a following process is required with respect to code each bit through the convolution coding.
That is, each bit is coded by a bit-in-word reading step reading x(n) in a current bit as a currently inputted single bit; a data shifting step performing a shifting operation with previously inputted data; an exclusive-logic operating step operating a 2-input exclusive logic with respect to each bit value included in selected cells of the shift register; and a storing step storing G
0
, G
1
which are the results from the previous step in a memory location, the sequence of the process being accomplished using the instructions of the DSP. Accordingly, since every step of the above process must be performed with respect to the entire data, it can be seen that the four steps must be carried out K×N times.
Since each step must be designated in a software, there are required a plurality of instructions. Further, although an assembly instruction is in charge of a single process and the process can be performed in one cycle, at least 10 cycles are needed since there are 5 shift register cells. Thus, there is required an operation volume which needs cycles of at least 10×N×K with respect to the entire input data.
In a commonly used digital signal processor, because an operation by each bit is needed to process the convolution coding, instructions performed by each bit are used but the process of the bit unit operation is not possibly performed in the DSP having a data bus system such as 16, 24 or 32 bits, for example. Although it is possible to process the bit unit operation, the bit unit operation must be repeatedly applied to a data word of a data bus width, that is, a 16, 24, or 32-bit width, as many as the number thereof. The convolution coding process consists of the shifting and the exclusive logic operation appropriate to the process by each bit unit, which are relatively simple operations. However, using read/write, shift, and exclusive logic operation instructions which are to be processed in a word unit of the DSP suitable to process a broad data bus width is not desirable since it can not be effectively utilized in usefulness of resources which are supplied for enabling the above process.
SUMMARY OF THE INVENTION
Accordingly, the present invention is directed to a convolution code generator which obviates the problems and disadvantages in the conventional art.
An object of the present invention is to provide a convolution code generator that is implemented in a hardware which is suitable for performing a convolution coding operation of a plurality of bits consisting of input data.
Another object of the present invention is to provide a convolution code generator appropriate for processing a convolution coding by a data bus width unit ill a digital signal processor which is suitable to process a relatively broad data bus width.
To achieve these and other advantages and in accordance with the purpose of the present invention, as embodied and broadly described herein, there is provided a convolution code generator, which generates an (n, k, m) convolution code, wherein n is a code word length, k is a message block length, and m is a memory block, which includes: a temporary data storing means for temporarily storing input data of a word of a predetermined data width consisting o

Affiliated with

Also associated with

Rating

Convolution code generator and digital signal processor... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Convolution code generator and digital signal processor..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Convolution code generator and digital signal processor... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2912548