Soft trellis decoding

• US 5,144,644 A
• Filed: 11/29/1989
• Issued: 09/01/1992
• Est. Priority Date: 10/13/1989
• Status: Expired due to Term

• Current Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Borth, David E.
• Primary Examiner(s): Safourek, Benedict V.
• Application Number:
  US07/442,971
• Time in Patent Office: 1,007 Days
• Field of Search: 371/43, 371/46, 371/44, 371/45, 375/14, 375/18, 375/34, 375/58, 375/96, 375/99, 375/100, 375/103, 375/94, 375/101, 364/724.11
• US Class Current: 375/341
• CPC Class Codes: H04B 7/005 Control of transmission; Equalising
• CPC Class Codes: H04L 25/03318 Provision of soft decisions
• CPC Class Codes: H04L 25/067 providing soft decisions, i.e. decisions together with an estimate of reliability H04L25/068 and H04L25/069 take precedence; sequence estimation techniques H04L25/03178
• CPC Class Codes: H04L 27/01 Equalisers baseband equalizers at the transmitter end H04L25/03343; in analogue transmission systems H04B3/04, H04B7/005

Abstract

There is provided a mechanism for soft trellis decoding. It comprises: generating decisions that rely upon other decision making and associating with it information indicative of the confidence held in those decisions and generated from the relative merit of any given decision. It further comprises: generating Maximum Likelihood Sequence Estimation equalization decisions and multiplying a signed value of the generated decision by the absolute values of the difference between alternative criterion utilized to arrive at any given decision, whereby the relative merit comprises a measure of signal quality at the time that given decision is made.

Claims

1. A method of soft trellis decoding comprising:

determining the probable present state from sequences of probable prior states and relative measures of merit related to the transitions from the last previous probable states to the probable present state

associating with the present state that measure of relative merit

and decoding the information represented by that more probable sequence of states utilizing the measure of merit associated with each state.

2. A method as claimed in claim 1 above, wherein such probable states are generated from Maximum-Likelihood Sequence Estimation equalizers.

3. A method as claimed in claim 1 above, wherein such probable states are generated from Viterbi algorithm equalizers.

4. A method as claimed in claim 1 above, wherein such decoding is generated from convolutional decoders.

5. A method as claimed in claim 1 above, wherein association comprises multiplying the value representative of the determined state by the measure of relative merit for that state.

6. A method as claimed in claim 5 above, wherein multiplication comprises multiplying a signed value representative of the determined state by the absolute value of the difference between alternative transition sequences.

7. A method as claimed in claim 1 above, wherein association comprises multiplying the value representative of the determined state by the absolute value of that measure of relative merit in the difference between alternative transition sequences.

8. A method as claimed in claim 1 above, wherein the relative merit comprises a measure of signal quality at the time that the present state is determined.

9. A method as claimed in claim 1 above, wherein the relative merit comprises the difference between alternative transition sequences.

10. Apparatus for soft trellis decoding comprising, operatively coupled in series:

means for determining the probable present state from sequences of probable prior states and relative measures of merit related to the transitions from the last previous probable states to the probable present state,

means for associating with the present state that measure of relative merit

and means for decoding the information represented by that more probable sequence of states utilizing the measure of merit associated with each state.

11. Apparatus for soft trellis decoding comprising, operatively coupled in series:

means for determining the probable present state from sequences of probable prior states and relative measures of merit related to the transitions from the last previous probable states to the probable present state,

means for multiplying the value representative of the determined state by the absolute value of that measure of relative merit in the difference between alternative transition sequences

and means for decoding the information represented by that more probable sequence of states utilizing the measure of merit weighted with each state.