AUTOMATIC FREQUENCY CONTROL UNDER LOW SIGNAL-TO-NOISE CONDITIONS

• US 20120163426A1
• Filed: 12/20/2011
• Published: 06/28/2012
• Est. Priority Date: 12/23/2010
• Status: Active Grant

• Current Assignee: Microchip Technology Incorporated
• Original Assignee: Microchip Technology Incorporated
• Inventors: Németh, József G., Kovács, Péter Szilveszter
• Granted Patent: US 8,548,033 B2
• Time in Patent Office: Days
• US Class Current: 375/219
• CPC Class Codes: H04L 2027/0046 Open loops
• CPC Class Codes: H04L 2027/0065 Frequency error detectors H04L2027/0067 takes precedence
• CPC Class Codes: H04L 2027/0071 Control of loops
• CPC Class Codes: H04L 2027/0095 in a preamble or similar structure
• CPC Class Codes: H04L 27/0014 Carrier regulation of chaotic carriers H04L27/001; for multicarrier receivers H04L27/2657

Abstract

Carrier frequency offset (CFO) is determined by sample-to-sample phase shifts of a digital radio signal. The CFO range is divided into a number of intervals and creates as many parallel derived streams as there are interval endpoints by pre-compensating (“back rotating”) the input by the sample-to-sample phase shift corresponding to the particular endpoint. Magnitude and phase values are computed of the correlation of a preamble pattern period with the preamble segment of each derived stream in parallel. The largest resulting magnitude value(s) are used to zoom in on the actual CFO present in the input stream. Improved accuracy in the presence of noise may be obtained by repeating the search for a shorter interval centered on the prior CFO value. Final CFO phase values from corresponding correlation computations then determine the actual CFO and corresponding sample-to-sample phase shift to be applied for pre-compensation (“back rotation”) in an open-loop AFC

Claims

1. A method for determining a carrier frequency offset and a corresponding sample-to-sample phase shift present in a sampled digital signal from a radio of a wireless transceiver, said method comprising the steps of:

a) uniformly dividing a carrier frequency offset (CFO) range [−

Δ

Ω

, Δ

Ω

) to be compensated, into ‘

N’

intervals of length, such that Δ

ω

=2·

Δ

Ω

/N<

δ

/(T·

L) where δ

<

1, T is the sampling period, L is the length of the preamble pattern in samples and the inequality sets the lower bound for the value of N, wherein δ

is the normalized length of the “

feasible range”

for direct phase calculation;

b) selecting midpoint values, Δ

ω

₀ Δ

ω

₁ . . . Δ

ω

<sub>N−

2</sub> Δ

ω

<sub>N−

1</sub>, of the intervals determined in step a);

c) connecting a stream of samples of an I-Q format input signal . . . s_i s<sub>i+L−

1</sub> simultaneously to N “

rotate &

correlate”

units numbered from 0 to N−

1 and assigning each midpoint value computed in step b) to the like numbered rotate &

correlate unit;

d) connecting an A output of the N rotate &

correlate units sequentially to the inputs of the “

smooth &

select”

unit;

e) starting the rotate &

correlate units on an external trigger that indicates a preamble pattern boundary in the input signal stream;

f) running a complete rotate &

correlate cycle while the rotate &

correlate units consume L input samples and compute the correlation vectors A_i=a·

e<sup>jφ

i</sup> in polar coordinates

2. A method for determining a carrier frequency offset and a corresponding sample-to-sample phase shift present in a sampled digital signal from a radio of a wireless transceiver, said method comprising the steps of:

dividing a carrier frequency offset (CFO) range to be covered into a plurality of intervals;

creating from a received signal as many parallel derived streams as there are intervals and pre-compensating (“

back rotating”

) signals received by the sample-to-sample phase shift corresponding to a midpoint of a particular one of the plurality of intervals;

computing magnitude and phase resulting from a correlation of an expected preamble pattern period waveform with a preamble segment of each derived stream in parallel;

applying if necessary, curve fitting (filtering) to resulting magnitude values in order to minimize noise effects;

selecting a largest resulting magnitude value(s) to zoom in on an actual CFO present in an input stream of the signals received;

repeating a search for a shorter interval centered on a CFO value located in a first run, if needed, in order to improve accuracy in the presence of noise and provided there is still input preamble left to work on; and

determining an actual CFO from the CFO values belonging to the selected interval and the result of the corresponding computed correlation phase.

3. A method for automatic frequency offset compensation under low signal-to-noise conditions, comprising the steps of:

a) detecting a preamble pattern boundary;

b) initiating an automatic frequency control (AFC) coarse step once the preamble pattern boundary has been detected;

c) determining whether the initiated AFC coarse step is a third AFC coarse step, wherein

if the initiated AFC coarse step is not the third AFC coarse step then going to step d), and

if the initiated AFC coarse step is the third AFC coarse step then going to step i);

d) taking a plurality of signal samples over a plurality of offset interval ranges;

e) averaging together the plurality of signal samples for each respective one of the plurality of offset interval ranges;

f) determining which one of the plurality of offset interval ranges has the averaged signal sample with a largest magnitude;

g) determining whether the largest magnitude is greater than an accept threshold, wherein

if the largest magnitude is greater than the accept threshold then going to step j), and

if the largest magnitude is not greater than the accept threshold then going to step h);

h) determining whether the largest magnitude is greater than a reject threshold, wherein

if the largest magnitude is greater than the reject threshold then returning to step a), and

if the largest magnitude is not greater than the reject threshold then going to step i);

i) rejecting the one of the plurality of offset interval ranges having the largest magnitude that is not greater than the reject threshold;

j) selecting a portion of the one of the plurality of offset interval ranges having the largest magnitude;

k) determining a first refined carrier offset estimate from the plurality of signal samples within the selected portion of the one of the plurality of offset interval ranges having the largest magnitude;

l) determining a second refined carrier offset estimate from the plurality of signal samples within the selected portion of the one of the plurality of offset interval ranges having the largest magnitude;

m) determining whether the AFC coarse step was a first or a second AFC coarse step, wherein

if the AFC coarse step was the second AFC coarse step than going to step n), and

if the AFC coarse step was the first AFC coarse step than going to step o);

n) averaging the first and second refined carrier offset estimates to provide an averaged carrier offset estimate then going to step q);

o) determining a third refined carrier offset estimate from the plurality of signal samples within the selected portion of the one of the plurality of offset interval ranges having the largest magnitude;

p) averaging the first, second and third refined carrier offset estimates to provide an averaged carrier offset estimate then going to step q); and

q) compensating a frequency offset from a carrier frequency with the averaged carrier offset estimate.

4. The method according to claim 3, wherein the plurality of offset interval ranges covers at least plus and minus 120 parts per million of the carrier frequency.

5. The method according to claim 3, wherein the each of the plurality of offset interval ranges comprises at least five of the plurality of signal samples.

6. The method according to claim 3, wherein the portion of the selected one of the plurality of offset interval ranges having the largest magnitude covers at least plus and minus 35 parts per million range of the selected one of the plurality of offset interval ranges representing a coarse carrier frequency offset of the carrier frequency.

7. The method according to claim 3, wherein the step of compensating the carrier frequency offset is done by adjusting a receiver local oscillator frequency.

8. The method according to claim 3, wherein the step of compensating the carrier frequency offset is done by de-rotating a raw I/Q waveform of received data symbols.

9. The method according to claim 3, wherein the step of determining which one of the plurality of offset interval ranges has the averaged signal sample with the largest magnitude uses sixteen (16) resources.

10. The method according to claim 9, wherein the sixteen (16) resources comprise sixteen (16) rotations of sixteen (16) possible reference symbols covering at least a 240 part per million frequency range.

11. The method according to claim 3, wherein the step of determining the first refined carrier offset estimate uses eight (8) resources.

12. The method according to claim 11, further comprising the step of determining another first refined carrier offset estimate using signal samples at Δ

t=+1 of the plurality of signal samples within the selected portion of the one of the plurality of offset interval ranges having the largest magnitude.

13. The method according to claim 12, wherein the step of determining the another first refined carrier offset estimate uses an additional eight (8) resources.

14. The method according to claim 3, wherein the step of determining the second refined carrier offset estimate uses eight (8) resources.

15. The method according to claim 14, further comprising the step of determining another second refined carrier offset estimate using signal samples at Δ

t=+1 of the plurality of signal samples within the selected portion of the one of the plurality of offset interval ranges having the largest magnitude.

16. The method according to claim 15, wherein the step of determining the another second refined carrier offset estimate uses an additional eight (8) resources.

17. The method according to claim 3, wherein the step of determining the third refined carrier offset estimate uses eight (8) resources.

18. The method according to claim 17, further comprising the step of determining another third refined carrier offset estimate using signal samples at Δ

t=+1 of the plurality of signal samples within the selected portion of the one of the plurality of offset interval ranges having the largest magnitude.

19. The method according to claim 18, wherein the step of determining the another third refined carrier offset estimate uses an additional eight (8) resources.

20. A method for automatic frequency offset compensation under low signal-to-noise conditions in a zero-intermediate frequency receiver, said method comprising the steps of:

receiving data signals having a preamble;

converting the received data signals into in-phase (I) and quadrature phase (Q) data signals at a zero intermediate frequency (IF);

sampling the I and Q data signals at the zero IF;

converting the sampled I and Q data signals to digital representations thereof with I and Q analog-to-digital converters (ADCs);

performing automatic frequency offset compensation of the received data signals, comprising the steps of;

a) detecting a preamble pattern boundary;

b) initiating an automatic frequency control (AFC) coarse step once the preamble pattern boundary has been detected;

c) determining whether the initiated AFC coarse step is a third AFC coarse step, wherein

if the initiated AFC coarse step is not the third AFC coarse step then going to step d), and

if the initiated AFC coarse step is the third AFC coarse step then going to step h);

d) averaging together the digital representations of the sampled I and Q data signals within each respective one of a plurality of offset interval ranges;

e) determining which one of the plurality of offset interval ranges has the averaged digital representations of the sampled I and Q data signals with a largest magnitude;

f) determining whether the largest magnitude is greater than an accept threshold, wherein

if the largest magnitude is greater than the accept threshold then going to step i), and

if the largest magnitude is not greater than the accept threshold then going to step g);

g) determining whether the largest magnitude is greater than a reject threshold, wherein

if the largest magnitude is greater than the reject threshold then returning to step a), and

if the largest magnitude is not greater than the reject threshold then going to step h);

h) rejecting the one of the plurality of offset interval ranges having the largest magnitude that is not greater than the reject threshold;

i) selecting a portion of the one of the plurality of offset interval ranges having the largest magnitude;

j) determining a first refined carrier offset estimate from the digital representations of the sampled I and Q data signals within the selected portion of the one of the plurality of offset interval ranges having the largest magnitude;

k) determining a second refined carrier offset estimate from the digital representations of the sampled I and Q data signals within the selected portion of the one of the plurality of offset interval ranges having the largest magnitude;

l) determining whether the AFC coarse step was a first or a second AFC coarse step, wherein

if the AFC coarse step was the second AFC coarse step than going to step m), and

if the AFC coarse step was the first AFC coarse step than going to step n);

m) averaging the first and second refined carrier offset estimates to provide an averaged carrier offset estimate then going to step p);

n) determining a third refined carrier offset estimate from the plurality of signal samples within the selected portion of the one of the plurality of offset interval ranges having the largest magnitude;

o) averaging the first, second and third refined carrier offset estimates to provide an averaged carrier offset estimate then going to step p); and

p) compensating a frequency offset from a carrier frequency of the received data signals with the averaged carrier offset estimate.

21. The method according to claim 20, further comprising the step of bandpass filtering the I and Q data signals at the zero IF.

22. The method according to claim 20, further comprising the step of decoding the digital representations of the I and Q data signals into data symbols.