Method and arrangement of estimating data sequences transmitted using viterbi algorithm

• US 5,450,445 A
• Filed: 11/02/1992
• Issued: 09/12/1995
• Est. Priority Date: 10/31/1991
• Status: Expired due to Fees

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Ushirokawa, Akihisa
• Primary Examiner(s): Chin, Stephen
• Assistant Examiner(s): VO, DON NGUYEN
• Application Number:
  US07/970,494
• Time in Patent Office: 1,044 Days
• Field of Search: 375/39, 375/80, 375/94, 375/99, 375/102, 371/43, 364/724.12
• US Class Current: 375/324
• CPC Class Codes: H04L 25/03184 Details concerning the metric

Abstract

A method and apparatus of estimating a data sequence transmitted using a Viterbi algorithm, are disclosed. Sampled values of incoming data are acquired at a predetermined time interval into a memory. Subsequently, a plurality of sampled values are retrieved from the memory. Further, a plurality of receive signal replicas, which have been previously prepared and stored in another memory, are retrieved. A plurality of receive signal replicas are determined by multiplying the sampled values by the receive signal replica coefficients. Branch metrics are then determined using the sampled values of the incoming data and the receive signal replicas. A data sequence is estimated using the branch metrics according to the Viterbi algorithm.

Claims

1. A method of estimating a data sequence transmitted using a Viterbi algorithm, comprising the steps of:

(a) storing sampled values of incoming data at a predetermined time interval in a first memory;

(b) retrieving a plurality of sampled values from said first memory;

(c) retrieving a plurality of receive signal replica coefficients previously prepared and stored in a second memory;

(d) determining a plurality of receive signal replicas by multiplying said plurality of sampled values retrieved in step (b) by said plurality of receive signal replica coefficients retrieved in step (c);

(e) determining branch metrics using said sampled values of the incoming data and said plurality of receive signal replicas determined in step (d); and

(f) estimating a data sequence using said branch metrics according to the Viterbi algorithm.

2. A method as claimed in claim 1, said plurality of sampled values and said plurality of receive signal replica coefficients are retrieved using survivor paths obtained using the Viterbi algorithm.

3. An apparatus for estimating a data sequence transmitted using a Viterbi algorithm, comprising:

a first memory (30) for storing a plurality of sampled values of incoming data at a predetermined time interval;

a second memory (32) for storing a plurality of receive signal replica coefficients which are used to determine receive signal replicas of all transitions branched from each state of a trellis diagram at a given time point;

a first controller (38) which receives said plurality of sampled values from said first memory and said plurality of receive signal replica coefficients from said second memory, in connection with a plurality of first set of states of a trellis diagram, said first controller producing a plurality of first pairs each consisting of a sampled value and a corresponding receive signal replica coefficient;

a second controller (40) which searches for, using history of survivor paths defined by a Viterbi algorithm, a plurality of sampled values from said first memory and a plurality of receive signal replica coefficients from said second memory both in connection with a plurality of second set of states of said trellis diagram, said second controller producing a plurality of second pairs each consisting of a sampled value and a corresponding receive signal replica coefficient;

a receive signal replica generator (42) which receives said plurality of first and second pairs and determines receive signal replicas in connection with said all transitions;

a branch metric calculator (34) which is coupled to receive said receive signal replicas and a plurality of sampled values of incoming data at said given time point and then determines branch metrics; and

a Viterbi processor (36) which is coupled to receive said branch metrics and estimates a data sequence, said Viterbi processor outputting said history of survivor paths.

4. A Viterbi estimator for estimating a data sequence transmitted using a trellis diagram including reduced states, comprising:

a first memory (50) for storing sampled values of incoming data at a predetermined time interval;

a second memory (52) for storing a plurality of receive signal replica coefficients which are used to determine receive signal replicas resulting from all transitions branched from trellis diagram'"'"'s reduced states at a given time point;

a replica calculation controller (58) which, when said replica calculation controller determines a non-singular state corresponding to a reduced state using history of survivor paths defined by a Viterbi algorithm, searches for a plurality of sampled values and a plurality of receive signal replica coefficients both related to the reduced state detected as a non-singular state, said replica calculation controller, when said replica calculation controller determines a singular state corresponding to a reduced state using the history of survivor paths, detecting a latest non-singular state on each of survivor paths defined by said Viterbi algorithm and then searching for a plurality of sampled values and a plurality of receive signal replica coefficients both related to the latest non-singular state detected, said replica calculation controller generating sampled values and receive signal replica coefficients which have been searched;

a receive signal replica generator (60) which receives said sampled values and said receive signal replica coefficients which are generated from said replica calculation controller, said receive signal replica generator determining receive signal replicas in connection with said all transitions of said trellis diagram including reduced states;

a branch metric calculator (54) which is coupled to receive said receive signal replicas and sampled values of incoming data and then determines branch metrics of said trellis diagram including reduced states; and

a Viterbi processor (56) which is coupled to receive said branch metrics and estimates a data sequence, said Viterbi processor outputting said history of survivor paths.