IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
US-0425030
(1995-04-19)
|
발명자
/ 주소 |
- Begin Guy (Verdun CAX) Thibeault Claude (St-Leonard CAX)
|
출원인 / 주소 |
- Universitedu Quebec aMontreal (Montreal CAX 03)
|
인용정보 |
피인용 횟수 :
7 인용 특허 :
0 |
초록
▼
The method is for correcting and decoding a sequence of branches representing encoded data bits into estimated information bits. The encoded data bits were previously encoded in a convolutional manner with v encoded symbols forming a branch where v is a predetermined value. The method comprising the
The method is for correcting and decoding a sequence of branches representing encoded data bits into estimated information bits. The encoded data bits were previously encoded in a convolutional manner with v encoded symbols forming a branch where v is a predetermined value. The method comprising the steps of (a) setting accumulation, correction and event indicators respectively into non-accumulating, non-active and incomplete states, and setting a precedent Nccb variable at a predetermined value N; (b) receiving the sequence of branches, for each of the branches received in the step (b) a series of verification and calculation being performed; and (c) verifying whether the branch corresponding to the oldest undelivered estimated information bit has gone through all the substeps of step (b), and if it is the case, delivering the oldest undelivered estimated information bit when the oldest undelivered estimated information bit is no longer convolutionally associated with any of the branches corresponding to the syndromes stored in the first and second registers. An apparatus for performing the method is also described.
대표청구항
▼
A method for correcting and decoding a sequence of branches representing encoded data bits into estimated information bits, the encoded data bits having been previously encoded in a convolutional manner with v encoded symbols forming a branch where v is a predetermined value, the method comprising t
A method for correcting and decoding a sequence of branches representing encoded data bits into estimated information bits, the encoded data bits having been previously encoded in a convolutional manner with v encoded symbols forming a branch where v is a predetermined value, the method comprising the steps of: a) setting accumulation, correction and event indicators respectively into non-accumulating, non-active and incomplete states, and setting a precedent Nccb variable at a predetermined value N; b) receiving the sequence of branches, for each of the branches received in the step (b): i) calculating (v-1) syndromes corresponding to the branch; ii) calculating an estimated information bit, storing the estimated information bit and calculating an actual Nccb variable based on the (v-1) syndromes and on the precedent Nccb variable; iii) verifying whether values of the precedent and actual Nccb variables are non-successive and whether the precedent Nccb variable is greater than or equal to N, and, if both conditions are positive, setting the accumulation indicator into an accumulating state and setting an Nss variable at zero; iv) verifying whether the accumulation indicator is in the accumulating state and whether the event indicator is in the incomplete state, and, if both conditions are positive, accumulating the (v-1) syndromes calculated in step (i) in a first register and incrementing the Nss variable by one; v) verifying whether the correction indicator is in an active state and whether the event indicator is in the incomplete state, and, if both conditions are positive, storing the (v-1) syndromes calculated in step (i) in a second register having a given length; vi) verifying whether the accumulation indicator is in the accumulating state, whether the precedent Nccb variable equals (N-1) and whether the actual Nccb variable equals N, and, if all conditions are positive: setting the accumulation indicator into the non-accumulating state, the correction indicator into an active state and the event indicator into a complete state; transferring all of the syndromes accumulated in the first register in the second register; and setting an Nmp variable at a value of the Nss variable; vii) verifying whether the value of the precedent Nccb variable equals (N-1), whether the value of the actual Nccb variable equals N, whether the accumulation indicator is in the non-accumulating state, whether the correction indicator is in the active state and whether the event indicator is in the incomplete state, and, if all conditions are positive, setting the event indicator into the complete state and setting the Nmp variable at a value representing the length of the second register; viii) verifying whether the accumulation indicator is in the accumulating state, whether the event indicator is in the incomplete state and whether the first register is full, and, if all conditions are positive: setting the accumulation indicator in the non-accumulating state and the correction indicator in the active state; transferring all of the syndromes accumulated in the first register to the second register; and setting the Nmp variable at a value representing the length of the second register; ix) verifying whether the correction indicator is in the active state, and if the correction indicator is in the active state: correcting estimated information bits convolutionally associated with the branch corresponding to the syndromes stored in the second register; updating the syndromes stored in the second register; verifying whether the event indicator is in the complete state and decrementing the Nmp variable by one if the event indicator is in the complete state; and verifying whether the Nmp variable equals zero and setting the event indicator into the incomplete state and the correction indicator into the non-active state if the Nmp variable equals zero; and c) verifying whether the branch corresponding to the oldest undelivered estimated information bit has gone through steps (i) to (ix), and if the branch corresponding to the oldest undelivered estimated information bit has gone through steps (i) to (ix), delivering the oldest undelivered estimated information bit when said oldest undelivered estimated information bit is no longer convolutionally associated with any of the branches corresponding to the syndromes stored in the first and second registers.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.