JPS59161741A - Error correcting code system - Google Patents

Abstract

PURPOSE:To determine the position of an error and to correct the error by setting two strings of check points and comparing codes of the check points with codes which are obtained newly from information digits according to parity check items. CONSTITUTION:Check points where a corrected check code change by an erroneous digit 13 are only a pair of H1 and S2. When the check code begins changing at H1, the change continues up to H2 along a check sequence (arrow). At this time, two kinds of check points exist on two check lines 5 and 6 respectively. When the check code beings changing at S2, the check codes change until S1 and two kinds of check points exist similarly. The erroneous digit and two kinds of check point exist on the same check line, so a closed chain is formed. Those check points form a chain with other check points of the same kind, and the check points discriminated in two kinds form the mutually independent chains with the erroneous digit, respectively. Those chains are used to determine the position of the erroneous digit, correcting the error.

Description

[Detailed Description of the Invention] (Techniques!i!T-) The present invention relates to a quasi-maximum 4 burst error correction code that can be constructed with a code length of Renci, a construction method thereof, and a correction method. .

(Background Art 9) Conventional burst error correction codes (shortened cyclic codes, fire codes, etc.) are constructed from a generator polynomial, and the error syndrome 't(4) is calculated by dividing the received code by the generator polynomial. All errors are corrected by cyclically shifting and adding it to the received code.Therefore, since the generator polynomial is added by 1XJi, it has the disadvantage that it is not possible to construct an arbitrary code length. Since correction is performed by calculation using a generator polynomial, the disadvantage is that the code structure and error correction are often sloppy.

(Problems to be solved by the invention) The present invention provides (1) a simple code construction method using two-dimensional parity transverse distortion points, and (2) all features of code construction using an arbitrary code length, and its purpose is to The purpose of this is to simplify the correction method, shorten the processing time, and miniaturize the device, which will be explained in detail below.

(Structure and operation of the invention) FIG. 1 shows an embodiment of the present invention, in which 1 is a tissue code, 2 is a transmission direction, 3 is a scanning direction H14 is a scanning direction S, 5 is an H inspection line, and 6 is the S inspection line, 7 is the inspection point, and 8 is the information point.

The systematic code 1 is composed of a square code array (nXn) that is transmitted sequentially along the transmission direction 2. Of the 40,000 times the two-dimensional plane is taken, the transmission is 10,000 m in two different directions (If
, S) 3.4. Pay 1 yuan to @1 [b? , t] is incorporated. At this time, a total of 2n test points 7 are included within the square code array. The amount to be sent is ¥1 (it can be incorporated into information point 8).

FIG. 2 shows one inspection point and two inspection lines (II, , 5') intersecting it. Suppose that arbitrary information is incorporated into all information points on the two inspection lines. At this time, if the result of calculating only the information points on each inspection line by modulo 2 is expressed as [I II, S ml, the following 4 traces [0, O], [0, I], [], , O ], C1゜]]
is possible. These 4 ways, i.e. 1 of 2 [bit:]
I resigned myself to 7 inches and had 1 [bit] of 1' information.
Prepare a total of 2 pod points, and calculate the result of 4 possible trajectories [0,0
].

[o, ]-], C1,,0], C], ,
]] against J, 6. Here, it is assumed that the inspection point 7 shown in FIG. 2 satisfies the parity inspection condition in the 1'1 inspection i 5 and does not satisfy the parity inspection condition in the S inspection; At this time, if another test point is placed on the S test line, test point 7 can satisfy the harness test condition on the S test line. On the other hand, if inspection point 7 satisfies the Harity test condition on the S inspection line but does not satisfy the H inspection condition, then if all the inspection points of 1' Oka on the H/risk line are placed, inspection point 7 will meet the H Inspection iJ: However, it is possible to satisfy the Harity inspection conditions.

Therefore, in order for the inspection point 7 to be a two-dimensional parity lateral distortion point, it is best to have at least one inspection point on each of the two inspection lines that intersect with it.

At the same time, the same applies to inspection points placed other than inspection point 7. For this reason, each test block has two test points CH2C5.

Give the value to each inspection point using the following procedure.

At the beginning, in the Yokokai Tsuzuri, the provisional footing of only the information points, the test code, is written as P"" X+ + I2 + ... + In-2 (n2
oti 2 ) given by (1). Here, I
: Information point.

Since each test 7 has two test points CII, C, and 5', their parity calibration equation becomes equation (2).

II+I2+"'+Inz+CjH+CB20 (m
oct2) (2) First, set CH1” on any one H test line x P found on that test line.Next, set C
CS0 on the S-test line that intersects with 1f is (2
) Substituting the previous CJl 1 into CH in the equation, 'l!' was found on this S check line. l constant parsimony code PS=I 1+ 12 +
...I,, 2 (tyvocl 2) f (
2) By substituting into the expression, it becomes difficult.

This is shown in FIG. Repeat this procedure until all check codes are reached.

Inspection series 9 is an arrangement of one test point with seven fixed test points. In general, an inspection series can consist of at least 4 inspection points, so one inspection I for 2n inspection points
In addition to the test sequence 1], a plurality of independent test sequences can be constructed.

When the sign of the information 7ton changes by one, there are two '
*Ashikenkura code changes. Here, if there is one test sequence, an arbitrary number of 'i7f report points change KZJ, and an even number of I codes always change within the transverse strain sequence. therefore,
The test code determined separately using method 2 from an arbitrary starting point matches the test code set at the last starting point. At this time, the two check lines that intersect with the starting points, the soma and the oyster, are satisfied at the same time, so all 2n parity transverse distortion equations are satisfied.

When all of the transverse strain edges constituting one inspection series are connected to each other by inspection lines, the inspection points become one closed chain on a two-dimensional plane. An example of a chain is shown in FIG.

(False signal detection and correction) The 7-drome obtained from the transmitted test points does not satisfy the harrity test condition for multiple test lines, so it is a true false signal.
In addition to 0, a vertical similar error code 11 is included.

Here, FIG. 5 shows -10 of the syndrome.

Each test is lonely, all the left sides of the parity test equation are tested,
When expressed as Q, it becomes equation (3).

Therefore, since all digino) acts as Takura Ko, is there a mistake in the inspection point? The selection is made so as not to contain any clutter. In FIG. 5, it is determined that there is no partial or total imposition of the shadow 12, and 2n 1ljl positive points are set in two columns in the transmission direction and the four directions [I5].

The code of the check point added for the Shungo is compared with the check code precisely determined according to the parity @ wild goose condition for the reception tee 7 notes of t from the frugal point, and the difference is found. Based on this difference, the check points are classified into two types: check points where the sign of the check point and @ sign match, and check points where the sign does not match.

Among the two test points in each test^9, 111m1
The 'a-constant transverse strain signal changes at the test point. The check code is a provisional fast food code that is sequentially modulo l/ll from an arbitrarily determined starting point.
], the test code changes at the test point where the Nakiashi test i code first changes. The next test code changes again at the test point where it changed, and the test code is 7QK horizontal bone.
- Matches the number. If the number of erroneous digits increases, it will also change to a positive request.

FIG. 6 shows the relationship between the erroneous digits 1-13 and the difference between the digits 1-13 and the test points. The ox arrow: indicates the taste of the appetizer.

According to Section 13, the test points where the general test code changes are (
There is only one set of H1, S2). When the check code starts to change at Hl, the change is along the transverse strain system Tali, H2-
1: The change continues. At this time, on each of the two @tube lines 5 and 6, there is one inspection point of each type. When the check code starts to change at S2, the check code changes at \S, -1, and similarly there are two types of check points. If you want, choose 11! Regardless of the starting point of the calibration series,
On the two horizontal distortion lines that intersect with the erroneous digit,
There are 11 inspection points for each of the two types of parts.

Since the erroneous digit and the test point of 2 points are on the same @ starvation, they each form a complete closed chain. The inspection point of the octopus is a part of the lateral strain 2Q, so it forms a hammer with other inspection points of the same bucket. Therefore, the test points determined as two knowledges form a closed knowledge that is independent of each erroneous digit.

Using this closed chain, the positions of all the tigits with errors are determined, and the errors are corrected. and check points, the code groove bottom is simpler and clearer than the shortened cyclic code or fire code, which consists of a generator polynomial. -! , there is an origin point that can be set with any code length. Since the inspection point is determined by the 7JII calculation of Method 2 according to the Harrity parsimony condition, there is an origin point that makes the side door processing easier. In the case of a square code array, uniform reversal is performed to maximize the minimum distance formation of each erroneous siidiji 71·I''Ji.

``Since the code is obtained by checking the code and corrects the error by forming the code, the processing time is reduced compared to the method of correcting the code by adding calculations, circular shifts, etc. to the received code. It is possible to shorten the processing time and downsize the processing equipment.

In the case of burst length b≦n'-1, the number of patterns captured on M-99 is CBVC, and the number of patterns shown by C6H-2 is large.

CB-Σ((n2-b-1-1)b2 Ct2)
(4) b=t Therefore, b≦n −] The correction for the burst p in the trench is 751 J. This is Gallage
≧2b is added to the condition N'- of the strict parsimony point a of r in all approximations. Here \N'=n, code length, k: number of information points. In other words, the organizational code in this paper is a sub-optimal parse!・It becomes a meditated correction code. Figure 1 shows the confusion of some chain codes and shortened J10J codes.

For the Zuda composite error pattern, find the number of song patterns and the number of chain patterns for the weight tK of the error pattern using the formula 1,
The value t that can be compared and corrected (necessary condition) was determined.

Figure 7 shows the results. It can be seen that the ability to correct the era name is also important when it comes to draft negotiations.

Table 1 (Effects of the invention) The present invention has a simple code structure and does not require all complicated arithmetic processing for error correction, so it has the advantage of shortening processing time and downsizing the processing equipment. It can be used primarily around digital information transmission terminals that require high-speed decoding performance and miniaturization.

[Brief explanation of the drawing]

Figure 1 shows one embodiment of the present invention, the tlr Asakari. Figure 2 shows the 11-grid affected point and the two transverse strain scars that intersect with it. Figure 3 shows the inspection point and the transverse strain series. Figure 4 is an example of silver produced by inspection points.
FIG. 5 is an explanatory diagram of the undrom's -1, FIG. 6 is an explanatory diagram of the relationship between one false city git and the fecal test point, and FIG. 7 is an explanatory diagram of the error correction ability and its evaluation. ]...Composition code (7x7), 2...Transmission direction, 3...
...Inspection direction H, 4・Inspection direction S, 5...H@Appetizer line,
6...S inspection line, 7...inspection point, 8...information point, 9
・Horizontal series, 10 ・True error, 11 ・Sakana error, 12 ・Shadow, 13-eD digit,] 4
...Harwning's upper bound, 15...Plot
The upper bound of kiB, 16...I'm disappointed. Patent Applicant Oki Electric Industry Co., Ltd. Patent Application Agent Megumi Yamamoto

Claims (1)

[Claims] In an error correction code system using an electronic processing device, it is a two-dimensional code array, and two inspections are performed in two calibration directions on a two-dimensional plane: two inspections each in the open, valley, and open directions. Arrange the points, the inspection points form one closed chain on the two-dimensional plane, and on the receiving side, determine 4f7i head loading without error from the syndrome,
Two rows of test points were set in the same direction as one transmission direction to ensure that there were no errors in the test points, and the test points were precisely determined according to the code of the set test points and the information digitization/harness test conditions. A correction code system characterized by distinguishing between true errors and pseudo errors by comparing the code with the code, determining the position of the error, and correcting the error.