Low bias method for estimating small signal-to-noise ratio
AssigneeInterDigital Technology Corporation
InventorsBin Li, Robert A. DiFazio, Ariela Zeira
Life of this patent
Estimating the term…
Abstract
A method for estimating signal-to-noise ratio (SNR) using a method with low bias that is effective for both positive SNRs and small to negative SNRs. The method is based on an iterative solution for the maximum likelihood estimate of the amplitude from which the SNR can be computed. The method is applicable for various modulated systems, including BPSK, QPSK and MPSK.
Claims18
A method for estimating signal-to-noise (SNR) ratio of a sequence of received BPSK communication symbols rk (k=1 to N) using a low bias method comprising:
- a) selecting an initial estimate of signal-to-noise ratio SNR0;
- b) calculating an initial amplitude value A0 and noise variance value σ0 based on the estimate SNR0;
- c) calculating a subsequent amplitude value A1 based on probability density;
- d) calculating a subsequent noise variance value σ1 based on amplitude value A1;
- e) calculating a subsequent signal-to-noise ratio SNR1 based on A1 and σ1;
- f) subtracting SNR0 from SNR1 to obtain a calculated resolution value;
- and
- g) adjusting SNR0 and repeating steps (b) through (f) until the calculated resolution value is equal to a predetermined acceptable value.
A method for estimating SNR of a sequence of received MPSK communication symbols rk (k=1 to N), with probability density rk=xk+jyk, using a low bias method comprising:
- a) selecting an initial estimate of signal-to-noise ratio SNR0;
- b) calculating an initial amplitude value A0 and noise variance value (σ0 based on the initial signal-to-noise value;
- c) calculating a subsequent amplitude value A1 using probability density;
- d) calculating a subsequent noise variance value σ1 based on amplitude value A1;
- e) calculating a subsequent signal-to-noise ratio SNR1 based on A1 and σ1;
- f) subtracting SNR0 from SNR1;
- g) adjusting SNR0 and repeating steps (b) through (f) until a predetermined resolution is achieved.
A method for estimating signal-to-noise ratio of a sequence of received BPSK communication symbols rk (k=1 to N) using a low bias method comprising:
- a) selecting minimum amplitude A0 and maximum amplitude A1 and an acceptable resolution Δ for iterative estimations of SNR;
- b) normalizing received symbols rk;
- c) calculating a mean Am of A0 and A1;
- d) calculating minimum noise variance σ0, maximum noise variance σ1, and mean noise variance σm;
- e) estimating amplitude values A′0, A′1, and A′m using probability density of the estimates Am of A0 and A1;
- f) updating A1=A′m if Am>A′m, else A0=A′m;
- g) deciding if A1−A2<Δ;
- h) setting final estimated amplitude AOUT equal to the mean of the final values of amplitudes A0 and A1 if step (g) is true, else returning to step (d) and repeating steps (c) through (g);
- and
- i) determining the final estimated SNR based on the estimated amplitude AOUT.
A method for estimating SNR of a sequence of received MPSK communication symbols rk (k=1 to N), with probability density rk=xk+jyk, using a low bias method comprising:
- a) selecting minimum amplitude A0 and maximum amplitude A1 and an acceptable resolution Δ for iterative estimations of SNR;
- b) normalizing received symbols rk;
- c) calculating a mean Am of A0 and A1;
- d) calculating minimum noise variance σ0, maximum noise variance σ1, and mean noise variance σm;
- e) estimating amplitude values A′0, A′1, and A′m using probability density of the estimates Am of A0 and A1;
- f) updating A1=A′m if Am>A′m, else A0=A′m;
- g) deciding if A1−A0<Δ is true;
- h) setting final estimated amplitude AOUT equal to the mean of the final values of amplitudes A0 and A1 if step (g) is true, else returning to step (d) and repeating the process;
- and
- i) determining the final estimated SNR based on the estimated amplitude AOUT.
A method for estimating SNR of a sequence of received MPSK communication symbols rk (k=1 to N), with probability density rk=xk+jyk, using a low bias method comprising:
- a) selecting minimum amplitude A0 and maximum amplitude A1 and an acceptable resolution Δ for iterative estimations of SNR;
- b) normalizing received symbols rk;
- c) calculating a mean Am of A0 and A1;
- d) calculating minimum noise variance σ0, maximum noise variance σ1, and mean noise variance σm;
- e) estimating amplitude values A′0, A′1, and A′m using probability density of the estimates Am of A0 and A1;
- f) updating A1=A′m if Am>A′m, else A0=A′m;
- g) repeating steps (d) through (f) for a predetermined number of iterations;
- h) setting final estimated amplitude AOUT equal to the mean of the final values of amplitudes A0 and A1;
- and
- i) determining the final estimated SNR based on the estimated amplitude AOUT.
Dependencies are inferred from the claim text.