US5319674A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Method for self-training adaptive equalization
Assignee: International Business Machines Corp
•
Filing: 1991-10-22
•
Granted: 1994-06-07
•
Expires: 2011-10-22
Abstract
Abstract A method is disclosed for adapting, in a self-training procedure, an equalizer based on distributed-arithmetic design. In such an equalizer, in each modulation interval addresses are derived from stored received signal samples x n q to choose some of the values stored in a look-up table. The chosen table values are combined by shift-and-add operations to obtain an equalizer output value y n . According to the novel adaptation scheme, only one of the chosen look-up table values is selected per modulation interval, for updating it in a particular fashion to reduce the error detected. The disclosed scheme ensures uniform selection of the stored look-up table values for updating. Only one additional shift-and-add operation is required per modulation interval to achieve satisfactory adaptation of the equalizer. No training sequence is required to be transmitted, and the scheme converges in all situations.
Inventors
- 👤 Giovanni Cherubini
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-10 01:31
CPC Classifications
H; H04; H04L; H04L25/00; H04L25/02; H04L25/03; H04L25/03006; H04L25/03012; H04L25/03019; H04L25/03038; H04L25/0305
Claims Analysis
Total Claims: 9
Claim 1
Independent Claim
248 words
A method of adapting a distributed-arithmetic equalizer for channel equalization in a partial response communication or storage system by updating table values d n (j) stored in a look-up table, said table values d n (j) for determining said equalizer characteristics, said equalizer generating an equalizer output value y n during each one of a plurality of successive modulation intervals, said method comprising the following steps executed during each successive modulation interval: fetching at least two table values d n (j) from said look-up table, the addresses of said fetched table values being defined by selected bits representing stored received signal samples x n q ; performing at least one shift-and-add operation on said fetched table values; defining a binary control value z n , said binary control signal represented by at least one predetermined bit z(O) selected from said stored signal samples x n q ; selecting, dependent on said binary control value, one of said fetched table values; generating a correction term -Δd n , said correction term being dependent on said equalizer output value y n and said binary control value z n ; adding said correction term -Δd n to said selected one d n of said fetched table values for providing an updated table value d n+1 ; and storing said updated table value d n+1 in said look-up table, said selected one d n of said fetched table values being replaced by said updated table value d n+1 in said look-up table.
Claim 2
Independent Claim
248 words
A method of adapting a distributed-arithmetic equalizer for channel equalization in a partial response communication or storage system by updating table values d n (j) stored in a look-up table, said table values d n (j) for determining said equalizer characteristics, said equalizer generating an equalizer output value y n during each one of a plurality of successive modulation intervals, said method comprising the following steps executed during each successive modulation interval: fetching at least two table values d n (j) from said look-up table, the addresses of said fetched table values being defined by selected bits representing stored received signal samples x n q ; performing at least one shift-and-add operation on said fetched table values; defining a binary control value z n , said binary control signal represented by at least one predetermined bit z(O) selected from said stored signal samples x n q ; selecting, dependent on said binary control value, one of said fetched table values; generating a correction term -Δd n , said correction term being dependent on said equalizer output value y n and said binary control value z n ; adding said correction term -Δd n to said selected one d n of said fetched table values for providing an updated table value d n+1 ; and storing said updated table value d n+1 in said look-up table, said selected one d n of said fetched table values being replaced by said updated table value d n+1 in said look-up table.
Claim 3
Dependent Claim
25 words
A method according to claim 1 including the additional step of limiting the value of said updated table value d n+1 between two predetermined values.
Claim 4
Dependent Claim
230 words
A method according to claim 1 wherein said step of generating said correction term -Δd n comprises the steps of: comparing said equalizer output value y n to a target set of symbols, said target set of symbols determined by the modulation scheme being used; choosing said correction term -Δd n equal to f 1 (z n ) if |y n | is greater than (M+γ), where f 1 (z n ) is a predetermined first function of said binary control value, M is the largest value in said target set of symbols, and γ is a first predetermined constant; choosing said correction term -Δd n equal to f 2 (z n )·x n-k .sbsb.0.sup.(B-z.sbsp.n) ·e n if |e n | is less than γ, where f 2 (z n ) is a predetermined second function of said binary control value, e n is the difference between said equalizer output value and the closest symbol in said target set of symbols, and x n-k .sbsb.0.sup.(B-z.sbsp.n) is a binary function of one selected bit of a predetermined one of the stored received signal samples x n q , selected dependent on whether said binary control value z n =1 or z n =0; and choosing said correction term -Δd n equal to f 3 (z n ) otherwise, where f 3 is a predetermined third function of said binary control value.
Claim 5
Dependent Claim
159 words
A method according to claim 1, including the additional steps of: storing N signal samples x n q each of said signal samples comprising (B+1) bits; defining (B+1) addresses for fetching said look-up table values, each from a group of N corresponding bits x n .sup.(w) for w=0, elipsis, B , by logically combining a predetermined companion bit (bc) of said N corresponding bits with each one of the remaining bits of that group; preselecting a given pair (VL-A, VL-B) from all said fetched table values and the pair of said predetermined companion bits (bc-A, bc-B) used for generating the addresses associated with said given pair of fetched table values, respectively; and dependent on said binary control value z n , selecting said one look-up table value d n of said given pair of fetched table values for updating and selecting its associated predetermined companion bit (bcs) as an additional input for generating said correction term -Δd n .
Claim 6
Dependent Claim
158 words
A method according to claim 3, including the additional steps of: storing N signal samples x n q , each of said signal samples comprising (B+1) bits; defining (B+1) addresses for fetching said look-up table values, each from a group of N corresponding bits x n .sup.(w) for w=0, elipsis, B, by logically combining a predetermined companion bit (bc) of said N corresponding bits with each one of the remaining bits of that group; preselecting a given pair (VL-A, VL-B) from all said fetched table values and the pair of said predetermined companion bits (bc-A, bc-B) used for generating the addresses associated with said given pair of fetched table values, respectively; and dependent on said binary control value z n , selecting said one look-up table value d n of said given pair of fetched table values for updating and selecting its associated predetermined companion bit (bcs) as an additional input for generating said correction term-Δd n .
Claim 7
Dependent Claim
27 words
A method according to claim 3, wherein said predetermined first function is defined as f.sub.1 (z.sub.n)=-2.sup.z.sbsp.n ·α.sub.1, where α 1 is a first predetermined power of two.
Claim 8
Dependent Claim
27 words
A method according to claim 3, wherein said predetermined second function is defined as f.sub.2 (z.sub.n)=-2.sup.-z.sbsp.n ·α.sub.2, where α 2 is a second predetermined power of two.
Claim 9
Dependent Claim
27 words
A method according to claim 3, wherein said predetermined third function is defined as f.sub.3 (z.sub.n)=2.sup.z.sbsp.n ·α.sub.3, where α 3 is a third predetermined power of two.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1997-06-06 | 1997-12-06 | 1998-06-07 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2001-06-06 | 2001-12-06 | 2002-06-07 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2005-06-06 | 2005-12-06 | 2006-06-07 | $7700.00 | Window Closed / Lapsed |