Frequency estimation method and system
AssigneeBenq Corporation
InventorsShih-Yang Yang, Zing-Wei Kang
Life of this patent
Estimating the term…
Abstract
A multiple linear regression frequency estimation method and system. A phase sequence of a received signal is derived and divided into several subsequences according to predetermined rules. The phase differences of the subsequences are calculated for frequency estimation using linear regression analysis. A plurality of frequency estimation values are obtained corresponding to each subsequence and combined.
Claims18
A multiple linear regression frequency estimation method, comprising:
- detecting a frequency correction burst (FB) signal carrying complex samples;
- calculating a phase sequence from the complex samples;
- dividing the phase sequence into a plurality of phase subsequences according to a predetermined rule;
- calculating phase errors of the phase sequence and each of the phase subsequences individually to obtain a phase error summation for the phase sequence and each of the phase subsequences;
- performing linear regression computation on each phase error summation to obtain linear regression slopes;
- and
- deriving a frequency from the linear regression slopes.
A multiple linear regression frequency estimation system, comprising:
- a detector, operating to detect a frequency correction burst (FB) signal carrying complex samples;
- a phase computing unit, operating to calculate a phase sequence from the complex samples received by the detector;
- a dividing unit, operating to divide the phase sequence into a plurality of phase subsequences according to a predetermined rule;
- a phase error computing unit, operating to calculate phase errors of the phase sequence and each of the phase subsequences individually and output a phase error summation for the phase sequence and each of the phase subsequences;
- a linear regression computing unit, operating to perform linear regression computation on each phase error summation generated by the phase error computing unit, in order to obtain linear regression slopes;
- and
- a frequency estimation unit, operating to derive a frequency from the linear regression slopes output from the linear regression computing unit.
A multiple linear regression frequency estimation method, comprising:
- detecting a frequency correction burst (FB) signal carrying complex samples;
- calculating a phase sequence from the complex samples;
- multiplexing the phase sequence evenly into at least two phase subsequences;
- calculating phase errors of each phase subsequence individually to obtain a phase error summation for each of the phase subsequences;
- performing linear regression computation on each phase error summation to obtain linear regression slopes;
- and
- deriving a frequency from the linear regression slopes.
A multiple linear regression frequency estimation system, comprising:
- a detector, operating to detect a frequency correction burst (FB) signal carrying complex samples;
- a phase computing unit, operating to calculate a phase sequence from the complex samples received by the detector;
- a multiplexer, operating to divide the phase sequence evenly into at least two phase subsequences;
- a phase error computing unit, operating to calculate phase errors of each of the phase subsequences individually and output a phase error summation for each of the phase subsequences;
- a linear regression computing unit, operating to perform linear regression computation on each phase error summation generated by the phase error computing unit, in order to obtain linear regression slopes;
- and
- a frequency estimation unit, operating to derive a frequency from the linear regression slopes output from the linear regression computing unit.
Dependencies are inferred from the claim text.