Back to results
US 2004/0071235 A1Published application
Low complexity high performance decoder and method of decoding for communications systems using multidimensional signaling
AssigneeTexas Instruments Incorporated
InventorsDavid J. Love, Srinath Hosur, Anuj Batra
Life of this patent
Estimating the term…
Abstract
A reduced search space minimum distance decoding algorithm provides average probability of error performance close to that of optimal MAP decoding. The decoding algorithm provides dramatic complexity reductions compared with MAP decoding.
Claims25
A method of decoding a signal vector, the method comprising the steps of:
- receiving signal vectors y1 . . . yk into a sub-optimal decoder and generating soft output bits therefrom;
- generating a reduced search space V via a reduced search space table creation unit in response to the soft output bits and an estimated channel H;
- and
- generating a signal vector ŝk via a maximum likelihood decoding unit in response to the reduced search space V and the signal vectors y1 . . . yk.
1 more element
A reduced search space decoder comprising:
- a sub-optimal decoder configured to receive a signal vector y1 . . . yk and generate soft output bits therefrom;
- a decision unit configured to generate a set E of error bit error patterns;
- a reduced search space table creation unit configured to generate a reduced search space V in response to the soft output bits and the indexes;
- and
- a minimum distance decoder configured to generate a multidimensional signal Ŝ in response to the reduced search space V and the signal vector y1 . . . yk.
2 more elements
A reduced search space decoder comprising:
- a reduced search space (RSS) decoder configured to receive symbol vectors y1 . . . yk and generate a hard symbol vector estimate {tilde over (S)} therefrom;
- and
- an interference cancellation decoder configured to generate a hard symbol vector Ŝ in response to the symbol vectors y1 . . . yk and the hard symbol vector estimate {tilde over (S)}.
A method of decoding a signal vector, the method comprising the steps of:
- receiving a symbol vector y1 . . . yk via a reduced search space maximum likelihood decoder and generating a hard symbol vector estimate S therefrom;
- and
- decoding the hard symbol vector estimate {tilde over (S)} via interference cancellation on each symbol and making a decision associated with the received symbol vector y1 . . . yk and the hard symbol vector estimate {tilde over (S)} to generate a symbol vector Ŝ.
Dependencies are inferred from the claim text.