SENSING WIRELESS COMMUNICATIONS IN TELEVISION FREQUENCY BANDS

• US 20110045781A1
• Filed: 08/17/2010
• Published: 02/24/2011
• Est. Priority Date: 08/18/2009
• Status: Active Grant

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Zhang, Wenyi, Sadek, Ahmed K., Shellhammer, Stephen J.
• Granted Patent: US 8,892,050 B2
• Time in Patent Office: Days
• US Class Current: 455/67.11
• CPC Class Codes: H04B 17/26 using historical data, averaging values or statistics
• CPC Class Codes: H04B 17/318 Received signal strength
• CPC Class Codes: H04B 17/382 for resource allocation, admission control or handover

Abstract

In general, techniques are described for sensing wireless communications in television frequency bands, which may be implemented by a sensing device comprising a sensing unit, a power spectral density (PSD) estimation unit, a filter unit, a candidate selection unit, an analysis unit and a decision unit. The sensing unit senses a signal in the television frequencies bands. The PSD estimation unit calculates an estimate of a PSD for the sensed signal. The filter unit filters the estimated PSD. The candidate selection unit analyzes the filtered PSD to identify a candidate frequency representative of a potentially in use frequency. The analysis unit computes a test statistic for the candidate frequency. The decision unit compares the test statistic to a threshold to identify whether the candidate frequencies is actively in use by wireless communication devices.

Claims

1. A method comprising:

sensing, with a sensing device, a signal in a range of frequencies reserved for use by television broadcast systems in broadcasting television content;

calculating, with the sensing device, an estimate of a power spectral density (PSD) for the sensed signal;

filtering, with the sensing device, the estimated PSD to remove noise and generate a filtered PSD for the sensed signal;

analyzing, with the sensing device, the filtered PSD to identifying one or more candidate frequencies within the range of frequencies, wherein the one or more candidate frequencies comprise frequencies potentially in use by one or more wireless communication devices;

computing, with the sensing device, one or more test statistics for each of the one or more candidate frequencies, wherein each of the one or more test statistics define a property of the one or more candidate frequencies; and

comparing, with the sensing device, the one or more test statistics to one or more thresholds to identify those of the candidate frequencies that are actively in use by the one or more wireless communication devices.

2. The method of claim 1, further comprising converting the estimated PSD from a linear domain to a logarithmic domain, wherein the logarithmic estimated PSD is expressed in a decibel format,

wherein filtering the estimated PSD comprises filtering the logarithmic estimated PSD to remove noise and generate the filtered PSD for the sensed signal.

3. The method of claim 1, wherein computing one or more test statistics comprises computing a height test statistic for at least one of the one or more candidate frequencies that defines a height property of the filtered PSD, wherein the height property is defined as a maximum value of the filtered PSD within a window of the filtered PSD minus a threshold value, wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

4. The method of claim 1, wherein computing one or more test statistics comprises computing a width test statistic for at least one of the one or more candidate frequencies that defines a width property of the filtered PSD, wherein the width property is defined as a difference between two values of the filtered PSD within a window of the filtered PSD where the filtered PSD intersects a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

5. The method of claim 1, further comprising converting the estimated PSD from a linear domain to a logarithmic domain, wherein the logarithmic estimated PSD is expressed in a decibel format,

wherein filtering the estimated PSD comprises filtering the logarithmic estimated PSD to remove noise and generate the filtered PSD for the sensed signal,

wherein computing one or more test statistics comprises computing an area test statistic for at least one of the one or more candidate frequencies that defines an area property of the filtered PSD, wherein the area property is defined as an area of the filtered PSD within a window of the filtered PSD as measured with respect to a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

6. The method of claim 1, wherein computing one or more test statistics comprises computing an area test statistic for at least one of the one or more candidate frequencies that defines an area property of the filtered PSD, wherein the area property is defined as an area of the filtered PSD within a window of the filtered PSD as measured with respect to a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

7. The method of claim 1, wherein computing one or more test statistics comprises computing a configurable test statistic in accordance with the following equation:

8. The method of claim 1, further comprising:

sensing, with the sensing device, multiple signals over time in a range of frequencies reserved for use by television broadcast systems in broadcasting television content;

calculating, with the sensing device, an estimate of a power spectral density (PSD) for each of the sensed signals;

filtering, with the sensing device, each of the estimated PSDs to remove noise and generate respective filtered PSDs for each of the sensed signals,

wherein computing one or more test statistics comprises;

forming a sequence of PSDs from the respective filtered PSDs;

computing a sequence of test statistics for the sequence of PSDs; and

calculating additional test statistics from the sequence of test statistics, wherein the additional test statistics include one or more of an average of the sequence of test statistics, a variance of the sequence of test statistics, a peak-to-peak variation of the sequence of test statistics, a generalized average of the sequence for test statistics and a correlation coefficient.

9. The method of claim 1, wherein analyzing the filtered PSD to identifying one or more candidate frequencies includes:

analyzing the filtered PSD to select a portion of the filtered PSD having a maximum value; and

identifying the center of the selected portion as a candidate frequency.

10. The method of claim 1, wherein analyzing the filtered PSD to identifying one or more candidate frequencies includes:

analyzing the filtered PSD to select a portion of the filtered PSD having a local maximum value; and

identifying the center of the selected portion as a candidate frequency.

11. The method of claim 1,

wherein the sensing device comprises a white space device, and

wherein the one or more wireless communication devices comprises at least one low-power wireless microphone, and

the method further comprising:

selecting, with the white space device, one or more of the candidate frequencies over which to communicate wirelessly based on the comparison.

40. The method of claim 1, further comprising instructions that cause the processor to compute a configurable test statistic in accordance with the following equation:

12. A sensing device for sensing wireless communications comprising:

a sensing unit that senses a signal in a range of frequencies reserved for use by television broadcast systems in broadcasting television content;

a power spectral density (PSD) estimation unit that calculates an estimate of a power spectral density (PSD) for the sensed signal;

at least one filter unit that filters the estimated PSD to remove noise and generate a filtered PSD for the sensed signal;

a candidate selection unit that analyzes the filtered PSD to identify one or more candidate frequencies within the range of frequencies, wherein the one or more candidate frequencies comprise frequencies potentially in use by one or more wireless communication devices;

one or more analysis units to compute one or more test statistics for each of the one or more candidate frequencies, wherein each of the one or more test statistics define a property of the one or more candidate frequencies; and

a decision unit that compares the one or more test statistics to one or more thresholds to identify those of the candidate frequencies that are actively in use by the one or more wireless communication devices.

13. The sensing device of claim 12, further comprising a decibel conversion unit that converts the estimated PSD from a linear domain to a logarithmic domain, wherein the logarithmic estimated PSD is expressed in a decibel format,

wherein the at least one filter unit filters the logarithmic estimated PSD to remove noise and generate the filtered PSD for the sensed signal.

14. The sensing device of claim 12, wherein the one or more analysis units compute a height test statistic for at least one of the one or more candidate frequencies that defines a height property of the filtered PSD, wherein the height property is defined as a maximum value of the filtered PSD within a window of the filtered PSD minus a threshold value, wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

15. The sensing device of claim 12, wherein the one or more analysis units compute a width test statistic for at least one of the one or more candidate frequencies that defines a width property of the filtered PSD, wherein the width property is defined as a difference between two values of the filtered PSD within a window of the filtered PSD where the filtered PSD intersects a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

16. The sensing device of claim 12, further comprising a decibel conversion unit that converts the estimated PSD from a linear domain to a logarithmic domain, wherein the logarithmic estimated PSD is expressed in a decibel format,

wherein the at least one filter unit filters the logarithmic estimated PSD to remove noise and generate the filtered PSD for the sensed signal,

wherein the one or more analysis units compute an area test statistic for at least one of the one or more candidate frequencies that defines an area property of the filtered PSD, wherein the area property is defined as an area of the filtered PSD within a window of the filtered PSD as measured with respect to a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

17. The sensing device of claim 12, wherein the one or more analysis units compute an area test statistic for at least one of the one or more candidate frequencies that defines an area property of the filtered PSD, wherein the area property is defined as an area of the filtered PSD within a window of the filtered PSD as measured with respect to a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

18. The sensing device of claim 12, wherein the one or more analysis units compute a configurable test statistic in accordance with the following equation:

19. The sensing device of claim 12,

wherein the sensing unit senses multiple signals over time in a range of frequencies reserved for use by television broadcast systems in broadcasting television content;

wherein the PSD estimation unit calculates an estimate of a power spectral density (PSD) for each of the sensed signals;

wherein the at least one filter unit filters each of the estimated PSDs to remove noise and generate respective filtered PSDs for each of the sensed signals,

wherein the one or more analysis units form a sequence of PSDs from the respective filtered PSDs, compute a sequence of test statistics for the sequence of PSDs, and calculate additional test statistics from the sequence of test statistics, wherein the additional test statistics include one or more of an average of the sequence of test statistics, a variance of the sequence of test statistics, a peak-to-peak variation of the sequence of test statistics, a generalized average of the sequence for test statistics and a correlation coefficient.

20. The sensing device of claim 12, wherein the candidate selection unit analyzes the filtered PSD to select a portion of the filtered PSD having a maximum value, and identifies the center of the selected portion as a candidate frequency.

21. The sensing device of claim 12, wherein the candidate selection unit further analyzes the filtered PSD to select a portion of the filtered PSD having a local maximum value and identifies the center of the selected portion as a candidate frequency.

22. The sensing device of claim 12,

wherein the sensing device comprises a white space device,

wherein the one or more wireless communication devices comprises at least one a low-power wireless microphone, and

wherein the decision unit further selects one or more of the candidate frequencies over which to communicate wirelessly based on the comparison.

23. An apparatus comprising:

means for sensing a signal in a range of frequencies reserved for use by television broadcast systems in broadcasting television content;

means for calculating an estimate of a power spectral density (PSD) for the sensed signal;

means for filtering the estimated PSD to remove noise and generate a filtered PSD for the sensed signal;

means for analyzing the filtered PSD to identifying one or more candidate frequencies within the range of frequencies, wherein the one or more candidate frequencies comprise frequencies potentially in use by one or more wireless communication devices;

means for computing one or more test statistics for each of the one or more candidate frequencies, wherein each of the one or more test statistics define a property of the one or more candidate frequencies;

means for comparing the one or more test statistics to one or more thresholds to identify those of the candidate frequencies that are actively in use by the one or more wireless communication devices.

24. The apparatus of claim 23, further comprising means for converting the estimated PSD from a linear domain to a logarithmic domain, wherein the logarithmic estimated PSD is expressed in a decibel format,

wherein the means for filtering the estimated PSD comprises means for filtering the logarithmic estimated PSD to remove noise and generate the filtered PSD for the sensed signal.

25. The apparatus of claim 23, wherein the means for computing one or more test statistics comprises means for computing a height test statistic for at least one of the one or more candidate frequencies that defines a height property of the filtered PSD, wherein the height property is defined as a maximum value of the filtered PSD within a window of the filtered PSD minus a threshold value, wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

26. The apparatus of claim 23, wherein the means for computing one or more test statistics comprises means for computing a width test statistic for at least one of the one or more candidate frequencies that defines a width property of the filtered PSD, wherein the width property is defined as a difference between two values of the filtered PSD within a window of the filtered PSD where the filtered PSD intersects a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

27. The apparatus of claim 23, further comprising means for converting the estimated PSD from a linear domain to a logarithmic domain, wherein the logarithmic estimated PSD is expressed in a decibel format,

wherein the means for filtering the estimated PSD comprises means for filtering the logarithmic estimated PSD to remove noise and generate the filtered PSD for the sensed signal,

wherein the means for computing one or more test statistics comprises means for computing an area test statistic for at least one of the one or more candidate frequencies that defines an area property of the filtered PSD, wherein the area property is defined as an area of the filtered PSD within a window of the filtered PSD as measured with respect to a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

28. The apparatus of claim 23, wherein the means for computing one or more test statistics comprises means for computing an area test statistic for at least one of the one or more candidate frequencies that defines an area property of the filtered PSD, wherein the area property is defined as an area of the filtered PSD within a window of the filtered PSD as measured with respect to a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

29. The apparatus of claim 23, wherein the means for computing one or more test statistics comprises means for computing a configurable test statistic in accordance with the following equation:

30. The apparatus of claim 23,

wherein the means for sensing comprise means for sensing multiple signals over time in a range of frequencies reserved for use by television broadcast systems in broadcasting television content;

wherein the means for calculating comprises means for calculating an estimate of a power spectral density (PSD) for each of the sensed signals;

wherein the means for filtering comprises means for filtering each of the estimated PSDs to remove noise and generate respective filtered PSDs for each of the sensed signals,

wherein the means for computing one or more test statistics comprises;

means for forming a sequence of PSDs from the respective filtered PSDs;

means for computing a sequence of test statistics for the sequence of PSDs; and

means for calculating additional test statistics from the sequence of test statistics, wherein the additional test statistics include one or more of an average of the sequence of test statistics, a variance of the sequence of test statistics, a peak-to-peak variation of the sequence of test statistics, a generalized average of the sequence for test statistics and a correlation coefficient.

31. The apparatus of claim 23, wherein the means for analyzing the filtered PSD to identifying one or more candidate frequencies includes:

means for analyzing the filtered PSD to select a portion of the filtered PSD having a maximum value; and

means for identifying the center of the selected portion as a candidate frequency.

32. The apparatus of claim 23, wherein the means for analyzing the filtered PSD to identifying one or more candidate frequencies includes:

means for analyzing the filtered PSD to select a portion of the filtered PSD having a local maximum value; and

means for identifying the center of the selected portion as a candidate frequency.

33. The apparatus of claim 23,

wherein the apparatus comprises a white space device,

wherein the one or more wireless communication device comprise at least one low-power wireless microphone, and

wherein the apparatus comprises means for selecting one or more of the candidate frequencies over which to communicate wirelessly based on the comparison.

34. A non-transitory computer-readable storage medium comprising instructions that cause a processor to:

sense a signal in a range of frequencies reserved for use by television broadcast systems in broadcasting television content;

calculate an estimate of a power spectral density (PSD) for the sensed signal;

filter the estimated PSD to remove noise and generate a filtered PSD for the sensed signal;

analyze the filtered PSD to identifying one or more candidate frequencies within the range of frequencies, wherein the one or more candidate frequencies comprise frequencies potentially in use by one or more wireless communication devices;

compute one or more test statistics for each of the one or more candidate frequencies, wherein each of the one or more test statistics define a property of the one or more candidate frequencies; and

compare the one or more test statistics to one or more thresholds to identify those of the candidate frequencies that are actively in use by the one or more wireless communication devices.

35. The non-transitory computer-readable storage medium of claim 34, further comprising instructions that cause the processor to:

convert the estimated PSD from a linear domain to a logarithmic domain, wherein the logarithmic estimated PSD is expressed in a decibel format; and

filter the logarithmic estimated PSD to remove noise and generate the filtered PSD for the sensed signal.

36. The non-transitory computer-readable storage medium of claim 34, further comprising instructions that cause the processor to compute a height test statistic for at least one of the one or more candidate frequencies that defines a height property of the filtered PSD, wherein the height property is defined as a maximum value of the filtered PSD within a window of the filtered PSD minus a threshold value, wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

37. The non-transitory computer-readable storage medium of claim 34, further comprising instructions that cause the processor to compute a width test statistic for at least one of the one or more candidate frequencies that defines a width property of the filtered PSD, wherein the width property is defined as a difference between two values of the filtered PSD within a window of the filtered PSD where the filtered PSD intersects a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

38. The non-transitory computer-readable storage medium of claim 34, further comprising instructions that cause the processor to:

convert the estimated PSD from a linear domain to a logarithmic domain, wherein the logarithmic estimated PSD is expressed in a decibel format;

filter the logarithmic estimated PSD to remove noise and generate the filtered PSD for the sensed signal; and

compute an area test statistic for at least one of the one or more candidate frequencies that defines an area property of the filtered PSD, wherein the area property is defined as an area of the filtered PSD within a window of the filtered PSD as measured with respect to a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

39. The non-transitory computer-readable storage medium of claim 34, further comprising instructions that cause the processor to compute an area test statistic for at least one of the one or more candidate frequencies that defines an area property of the filtered PSD, wherein the area property is defined as an area of the filtered PSD within a window of the filtered PSD as measured with respect to a threshold value, and wherein the threshold value identifies a median or a mean of all values of the filtered PSD within the window.

41. The non-transitory computer-readable storage medium of claim 34, further comprising instructions that cause the processor to:

sense multiple signals over time in a range of frequencies reserved for use by television broadcast systems in broadcasting television content;

calculate an estimate of a power spectral density (PSD) for each of the sensed signals;

filter each of the estimated PSDs to remove noise and generate respective filtered PSDs for each of the sensed signals;

form a sequence of PSDs from the respective filtered PSDs;

compute a sequence of test statistics for the sequence of PSDs; and

calculate additional test statistics from the sequence of test statistics, wherein the additional test statistics include one or more of an average of the sequence of test statistics, a variance of the sequence of test statistics, a peak-to-peak variation of the sequence of test statistics, a generalized average of the sequence for test statistics and a correlation coefficient.

42. The non-transitory computer-readable storage medium of claim 34, further comprising instructions that cause the processor to:

analyze the filtered PSD to select a portion of the filtered PSD having a maximum value; and

identify the center of the selected portion as a candidate frequency.

43. The non-transitory computer-readable storage medium of claim 34, further comprising instructions that cause the processor to:

analyze the filtered PSD to select a portion of the filtered PSD having a local maximum value; and

identify the center of the selected portion as a candidate frequency.

44. The non-transitory computer-readable storage medium of claim 34,

wherein the non-transitory computer-readable storage medium is included within a white space device,

wherein the one or more wireless communication device comprise at least one low-power wireless microphone, and

wherein the instructions further cause the processor to select one or more of the candidate frequencies over which to communicate wirelessly based on the comparison.

45. A method for obtaining features of a high-resolution power spectral density (PSD) of a detected signal, the method comprising:

determining a height of the center of an energy concentration having a main lobe, wherein the height is measured as a difference between S[kc] and a reference floor level, wherein the reference floor level captures any surrounding, non-energy concentration, and noise, and wherein the reference floor level is estimated as the mean, median, or minimum of S[k], k=−

K+1, . . . , −

1, 0, 1, . . . , K;

determining a width of the main lobe of the energy concentration, wherein the width is measured as a total number of frequency bins surrounding kc satisfying S[k]>

S[kc]/W, wherein W>

1 is a threshold parameter, and wherein the total number of frequency bins may be converted into hertz, and in order to refine the width estimate, including interpolating; and

determining a power of the main lobe of the energy concentration, wherein the power is measured as the ratio between the power of frequency bins surrounding kc, and a total power expressed as

46. The method of claim 45, further comprising detecting a wireless microphone in a whitespace environment by comparing the height, width, and power metrics therein with predefined thresholds.

47. A method of determining whether the detected signal is a legitimate wireless microphone signal, when a co-channel signal is an NTSC signal, the method comprising detecting a peak at a frequency equal to 5.75 MHz greater than the lower band edge frequency.

48. The method of claim 47, further comprising detecting a spurious signal caused by intermodulation of upper or lower adjacent NTSC signals.

49. The method of claim 47, further comprising comparing a frequency of a detected signal with an a-priori list of possible signal frequencies to determine whether the detected signal is legitimate.