T 1445/20 (Error overflow/HUAWEI) 27-04-2023
Download and more information:
Decoding method and decoding device
Decision in written proceedings - (yes): indication of non-attendance by applicant - oral proceedings neither necessary nor expedient
Admittance of request filed with the appeal - sole claim request (no): amendment not suitable to address the issues which led to the appealed decision
I. The appeal lies from the decision of the examining division to refuse the present European patent application. The sole claim request subject to the appealed decision was deemed not to be allowable under Articles 83 and 84 EPC.
II. The appellant requests that the decision under appeal be set aside and that a patent be granted according to the claims of an amended claim request filed for the first time with the statement setting out the grounds of appeal.
III. The appellant was summoned to oral proceedings before the board. A communication was issued under Article 15(1) RPBA 2020 including the board's negative preliminary opinion regarding the admittance into the proceedings under Article 12(4) RPBA 2020 of the amended claim request.
IV. In a written reply, the appellant stated that it would not be attending the arranged oral proceedings.
V. Subsequently, the oral proceedings were cancelled.
VI. Claim 1 of the amended claim request reads as follows
(board's feature labelling and highlighting of amendments vis-à-vis claim 1 of the sole claim request underlying the appealed decision):
(a) "An error-correcting decoding method, comprising:
(b) performing (S201) iterative decoding on a multidimensional code in order to obtain incorrigible code words, wherein the number of error bits of an incorrigible code word in each dimension where error bits of the incorrigible code words are located is greater than the [deleted: capacity ]error correction capability of the multidimensional code;
(c) determining (S202) locations of error bits in the incorrigible code words that have been obtained by performing the iterative decoding on the multidimensional code, wherein the locations of the error bits in the incorrigible code words are multidimensional coordinate locations of the error bits in the multidimensional code;
(d) correcting (S203) error bits of a part of the incorrigible code words in the multidimensional code according to the determined locations of the error bits in the incorrigible code words, so that the number of error bits of a code word in each dimension where the error bits of the incorrigible code words are located is less than the [deleted: capacity ]error correction capability of the multidimensional code, wherein only error bits that are located in both a first incorrigible code word and a second incorrigible code word are corrected; and
(e) after the error bits of a part of the incorrigible code words in the multidimensional code are corrected, performing (S204) iterative decoding on the multidimensional code".
1. Decision in written proceedings
As the appellant effectively withdrew its request for oral proceedings by declaring its intent not to attend them, and as the board does not consider the conduct of oral proceedings to be expedient either (cf. Article 116(1) EPC), the decision is handed down in written proceedings (Article 12(8) RPBA 2020).
2. Technical background
2.1 The present application relates to correcting, by means of an error-correction algorithm, errors that typically occur when a coded bit stream is transmitted in a communication system. In particular, the present application concerns "incorrigible code words" that are obtained by iterative decoding of such a coded bit stream. These "incorrigible code words" are inherent to any error-correction algorithm. They arise when the number of error bits in the coded bitstream is beyond the algorithm's "capability" to correct. In such a case, the error-correction algorithm is overloaded.
2.2 The invention addresses this overloading by first locating the error bits in the incorrigible code words and then correcting some of these error bits. This correction takes place until the number of error bits is below the error correction capability of the algorithm, after which iterative decoding is resumed.
3. Amended claim request: admittance into the proceedings
3.1 In view of the amendments made to present claim 1, the admittance into the proceedings of the amended claim request mentioned in point II above is at the board's discretion under Article 12(4) RPBA 2020.
3.2 The amended claim request is not suitable to address all of the issues which led to the decision under appeal. In particular, the objections raised in Reasons 1.3 to 1.5 and 2 of the appealed decision are unaffected by claim 1 of the amended claim request. The board notes the following in this respect:
3.2.1 In Reasons 1.3 of the appealed decision, the examining division correctly observed that the expression "in order to obtain incorrigible code words" according to feature (b) renders claim 1 of the appealed decision's claim request unclear. The board endorses the reasons provided by the examining division in this respect. These reasons still apply to claim 1 of the amended claim request. Even with "a mind willing to understand the claims", the skilled reader would not necessarily interpret, based on their common general knowledge, the "iterative decoding" of feature (b) as "a first round of iterative decoding" in the sense that only a single iteration of the iterative decoding would be conducted. Incidentally, this would also run contrary to the present application's description (see page 6, lines 16 and 17 and page 9, lines 14 to 24): the second and third embodiment's decoder is explicitly configured to perform the iterative decoding according to feature (b) "m times" (emphasis added).
3.2.2 The board also concurs with Reasons 1.4 and 1.5 of the impugned decision. It is indeed not clear how the "iterative decoding" of feature (b) can be distinguished from that of feature (e). Even a skilled reader "well versed in iterative coding" would not immediately understand, based on claim 1 taken by itself, that these features represent two (different) stages of iterative decoding, where the first stage would "identify the incorrigible words" and the second stage would then correct "the remaining parts of the words". Moreover, it is not apparent how the "determining" step of feature (c) and the "correcting" step according to feature (d) relate to the "iterative decoding" of features (b) and (e). In particular, it is unclear whether these steps are part of the "iterative decoding" or whether they represent additional steps.
3.2.3 As to Reasons 2 of the appealed decision, the board agrees with the examining division at least in that it is not sufficiently disclosed for the skilled person how the "correcting" according to feature (d) indeed takes place. The description (see e.g. page 5, lines 9 to 12 regarding step S103; page 7, line 34 to page 8, line 3 regarding step S203 and on page 11, lines 4 to 6 and 19 to 24 regarding error correcting unit 53) of the present application merely states that error bits of the incorrigible word are corrected but does not teach how this is done. The board notes in this respect that code words that are damaged beyond the multidimensional code's ability to repair will typically necessitate a re-transmission of the damaged data. The skilled person would not understand the "correcting" step of feature (d) to amount to a mere re-transmission. Moreover, the present application is silent about any such re-transmission.
3.3 The board therefore decided not to admit the amended claim request into the proceedings under Article 12(4) RPBA 2020.
4. Conclusion
Already because there is no admissible claim request on file, the appeal is to be dismissed.
Order
For these reasons it is decided that:
The appeal is dismissed.