Title:
【発明の名称】神経網による最尤シーケンス推定を用いたデジタル適応受信器
Document Type and Number:
Japanese Patent JP2809643
Kind Code:
B2
Abstract:
A maximum-likelihood sequence estimator receiver includes a matched filter (56) connected to a digital transmission channel (58) and a sampler (60) for providing sampled signals output by the matched filter (56). The sampled signals are input to an analog neural network (68) to provide high-speed outputs representative of the transmission channel signals. The neural network outputs (70) are also provided as inputs to a coefficient esitmator (78) which produces coefficients for feedback to the matched filter (56). For time-varying transmission channel characteristics, the coefficient estimator (78) provides a second coefficient output which is utilized for changing the interconnection strenghts of the neural network connection matrix to offset the varying transmission channel characteristics.
Inventors:
John Dee Provence
Application Number:
JP13536188A
Publication Date:
October 15, 1998
Filing Date:
June 01, 1988
Export Citation:
Assignee:
Texas Instruments Incorporated
International Classes:
H04L25/08; G06J1/00; G06N3/04; H03M13/23; H04B3/06; H04L25/03; (IPC1-7): H04L25/08; H03M13/12; H04L25/03
Domestic Patent References:
| JP63198131A | ||||
| JP62114071A |
Attorney, Agent or Firm:
Minoru Nakamura (4 outside)