Method and apparatus for recursive filtering of parallel intermittent streams of unequally reliable time discrete data
DCFirst Claim
1. A method for recursive filtering of time discrete measurement data relating to a measurable time-varying attribute to generate an estimated magnitude for the attribute at a particular time instant, comprising the steps of:
- collecting at least one measurement of the time-varying attribute during said time instant, wherein each said measurement has associated therewith a reliability indicator representing a relative reliability of said measurement;
filtering the reliability indicators for the at least one measurement to produce a filtered reliability indicator, said filtered reliability indicator including a time-weighted historical reliability indicator, wherein the historical reliability indicator includes information relating to a relative reliability of at least one measurement collected prior to said time instant;
filtering the at least one measurement to produce an estimated magnitude of said attribute for said time instant, said estimated magnitude including;
a time-weighted and reliability-weighted historical estimated magnitude, wherein the historical estimated magnitude includes information relating to at least one measurement of the time-varying attribute collected prior to said time instant; and
a reliability-weighted average of said at least one measurement, said reliability weighting of the reliability-weighted historical estimated magnitude and the reliability-weighted average of said at least one measurement performed using said filtered reliability indicator.
1 Assignment
Litigations
0 Petitions
Accused Products
Abstract
An apparatus and method for recursive filtering of time discrete measurement data relating to a measurable time-varying attribute to generate an estimated magnitude for the attribute at a particular time instant. Each measurement has an associated reliability indicator representing a relative reliability of the measurement. Using the individual reliability indicators and a time-weighted filtered reliability indicator for a prior time instant, a filtered reliability indicator for the current time instant is generated. The filtered reliability indicator for the current time instant is then used to reliability-weight a weighted average of the plurality of measurements and to reliability-weight a time-weighted estimated magnitude for the prior time instant. The weighted values are then used to generate an estimated magnitude of the time-varying attribute for the current time instant.
30 Citations
35 Claims
-
1. A method for recursive filtering of time discrete measurement data relating to a measurable time-varying attribute to generate an estimated magnitude for the attribute at a particular time instant, comprising the steps of:
-
collecting at least one measurement of the time-varying attribute during said time instant, wherein each said measurement has associated therewith a reliability indicator representing a relative reliability of said measurement;
filtering the reliability indicators for the at least one measurement to produce a filtered reliability indicator, said filtered reliability indicator including a time-weighted historical reliability indicator, wherein the historical reliability indicator includes information relating to a relative reliability of at least one measurement collected prior to said time instant;
filtering the at least one measurement to produce an estimated magnitude of said attribute for said time instant, said estimated magnitude including;
a time-weighted and reliability-weighted historical estimated magnitude, wherein the historical estimated magnitude includes information relating to at least one measurement of the time-varying attribute collected prior to said time instant; and
a reliability-weighted average of said at least one measurement, said reliability weighting of the reliability-weighted historical estimated magnitude and the reliability-weighted average of said at least one measurement performed using said filtered reliability indicator. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
collecting at least one other measurement of the time-varying attribute during the time instant, wherein each said at least one other measurement has associated therewith a reliability indicator representing a relative reliability of said measurement, said at least one other measurement comprising a second subset of the plurality of measurements;
filtering the reliability indicators for the at least one other measurement to produce a filtered reliability indicator associated with the second subset;
filtering the at least one other measurement to produce an estimated magnitude associated with the second subset; and
combining the estimated magnitude associated with the first subset with the estimated magnitude associated with the second subset to produce a larger subset estimated magnitude of the time-varying attribute for the time instant for a larger subset of the plurality of measurements, said larger subset including at least the first and second subsets of the plurality of measurements.
-
-
3. The method of claim 2, wherein the step of combining the first subset estimated magnitude with the second subset estimated magnitude to produce the larger subset estimated magnitude includes reliability-weighting the first subset estimated magnitude using the first subset filtered reliability indicator and reliability-weighting the second subset estimated magnitude using the second subset filtered reliability indicator.
-
4. The method of claim 3, wherein the reliability-weighting of the first subset estimated magnitude is performed in accordance with a ratio of the first subset reliability indicator to a reliability indicator for all of the plurality of measurements in the larger subset and the reliability-weighting of the second subset estimated magnitude is performed in accordance with a ratio of the second subset reliability indicator to the reliability indicator for all of the plurality of measurements in the larger subset.
-
5. The method of claim 2, wherein the larger subset comprises all of the plurality of measurements.
-
6. The method of claim 2, wherein at least one of the plurality of measurements comprises an assigned value indicating that no measurement data was collected for said at least one of the plurality of measurements.
-
7. The method of claim 1, wherein the at least one measurement comprises a plurality of measurements.
-
8. The method of claim 7, further comprising the step of determining the reliability-weighted average of said plurality of measurements, said step of determining comprising:
-
calculating the product of each measurement and the associated reliability indicator for each of the plurality of measurements; and
combining the calculated products for the plurality of measurements.
-
-
9. The method of claim 8, wherein the step of determining the reliability-weighted average of said plurality of measurements further comprises:
weighting the combined calculated products according to a ratio of the sum of the reliability indicators associated with the measurements to the filtered reliability indicator.
-
10. The method of claim 1, wherein the historical reliability indicator and the historical estimated quantity are time-weighted by a factor (1−
- β
), where β
comprises a preselected parameter that determines how fast weight of the historical reliability indicator and the historical estimated quantity decrease with time.
- β
-
11. The method of claim 10, wherein the reliability-weighted average of said at least one measurement is further weighted by the parameter β
- .
-
12. The method of claim 10, wherein the step of filtering the reliability indicators for the at least one measurement to produce a filtered reliability indicator comprises the step of combining the reliability indicators for the at least one measurement to form a combined reliability indicator, the filtered reliability indicator including the combined reliability indicator weighted by the parameter β
- .
-
13. The method of claim 1, wherein the at least one measurement of the time-varying attribute comprises at least one quality measurement of at least one signal in a telecommunications network.
-
14. The method of claim 13, wherein the at least one quality measurement indicates a quality of the at least one signal transmitted over an air interface in a mobile telecommunications network.
-
15. The method of claim 1, further comprising the step of determining the time-weighted and reliability-weighted historical estimated quantity, wherein said step of determining comprises weighting the historical estimated quantity according to a ratio of the historical reliability indicator to the filtered reliability indicator.
-
16. The method of claim 15, further comprising the step of determining the reliability-weighted average of said at least one measurement according to a ratio of the sum of the reliability indicators associated with the at least one measurement to the filtered reliability indicator.
-
17. The method of claim 1, wherein the historical estimated quantity comprises an estimated quantity of said attribute for a prior time instant.
-
18. The method of claim 1, wherein the at least one measurement of the time-varying attribute is collected by at least one sensor.
-
19. The method of claim 1, wherein the step of collecting at least one measurement of the time-varying attribute during said time instant comprises:
-
identifying at least one sampling instant within said time instant;
attempting to collect measurement data at each sampling instant;
assigning at least one measurement value for each sampling instant regardless of whether the attempt to collect measurement data succeeded;
assigning a zero value to the reliability indicator associated with each measurement for which the attempt to collect measurement data failed; and
assigning a non-zero value to the reliability indicator associated with each measurement for which the attempt to collect measurement data succeeded.
-
-
20. The method of claim 1, further comprising the step of determining the time-weighted and reliability-weighted historical estimated quantity, wherein said step of determining comprises weighting the historical estimated quantity according to:
-
where β
comprises a preselected parameter that determines how fast weight of the historical reliability indicator and the historical estimated quantity decrease with time, x comprises a sum of the reliability indicators associated with the measurements, and z comprises the filtered reliability indicator.
-
-
21. An apparatus for recursively filtering measurement data, comprising:
-
a buffer for storing a plurality of measurements of a time-varying attribute, wherein each said measurement has associated therewith a reliability indicator representing a relative reliability of said measurement, each of said measurements of the time-varying attribute made during a particular time instant;
a memory for storing filtering instructions, said filtering instructions used to generate an estimated magnitude for the time-varying attribute at the particular time instant; and
a processor for filtering the plurality of measurements in accordance with the filtering instructions, said processor operating to;
time-weight a historical estimated magnitude, wherein the historical estimated magnitude includes information relating to at least one measurement of the time-varying attribute collected prior to the particular time instant; and
generate a weighted average of the plurality of measurements using the reliability indicator associated with each of the plurality of measurements; and
generate an estimated magnitude of the time-varying attribute for the particular time instant using the weighted historical estimated magnitude and the weighted average of the plurality of measurements. - View Dependent Claims (22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32)
filter the reliability indicators associated with the plurality of measurements to produce a filtered reliability indicator, said filtered reliability indicator determined using a time-weighted historical reliability indicator, wherein the historical reliability indicator includes information relating to a relative reliability of at least one measurement collected prior to said time instant;
reliability-weight the time-weighted historical estimated magnitude and the weighted average of the plurality of measurements using the filtered reliability indicator; and
wherein the generation of the estimated magnitude of the time-varying attribute for the particular time instant includes using the reliability-weighted and time-weighted historical estimated magnitude and the reliability-weighted average of the plurality of measurements.
-
-
23. The apparatus of claim 22, wherein:
-
the weighted average of the plurality of measurements is reliability-weighted according to a ratio of a sum of the reliability indicators associated with the plurality of measurements to the filtered reliability indicator; and
the time-weighted historical estimated magnitude is reliability-weighted according to a ratio of the historical reliability indicator to the filtered reliability indicator.
-
-
24. The apparatus of claim 22, wherein:
-
the time-weighting of the historical reliability indicator includes time-weighting said historical reliability indicator by a factor (1−
β
) where β
comprises a preselected parameter that determines how fast weight of the historical reliability indicator and the historical estimated quantity decrease with time; and
said filtered reliability indicator is further determined using a combination of the reliability indicators associated with the plurality of measurements, said combination weighted by a factor β
.
-
-
25. The apparatus of claim 22, wherein the historical estimated quantity comprises an estimated quantity of the time-varying attribute for a prior time instant.
-
26. The apparatus of claim 22, wherein each of the plurality of measurements comprises a quality measurement of a signal in a telecommunications network.
-
27. The apparatus of claim 21, further comprising at least one sensor for collecting the plurality of measurements.
-
28. The apparatus of claim 27, wherein a quantity of measurements collected during a prior time instant differs from a quantity of measurements included in the plurality of measurements, wherein the processor further operates to further weight the historical estimated magnitude and the weighted average of the plurality of measurements to account for said difference in generating the estimated magnitude of the time-varying attribute for the particular time instant.
-
29. The apparatus of claim 27, wherein the at least one sensor operates to:
-
attempt to collect measurement data during at least one sampling instant of the particular time instant;
assign at least one measurement value for each sampling instant regardless of whether the attempt to collect measurement data succeeded;
assign a zero value to the reliability indicator associated with each measurement for which the attempt to collect measurement data failed; and
assign a non-zero value to the reliability indicator associated with each measurement for which the attempt to collect measurement data succeeded.
-
-
30. The apparatus of claim 21, wherein the plurality of measurements are unequally reliable.
-
31. The apparatus of claim 21, wherein the generation of the weighted average of the plurality of measurements includes:
-
calculating the product of each measurement and the reliability indicator associated with the measurement; and
combining the calculated products for the plurality of measurements.
-
-
32. The apparatus of claim 21, wherein the time-weighting of the historical estimated magnitude includes time-weighting said historical estimated magnitude by a factor (1−
- β
), where β
comprises a preselected parameter that determines how fast weight of the historical estimated quantity decreases with time, and wherein the generation of the weighted average of the plurality of measurements further includes time-weighting the weighted average of the plurality of measurements by a factor β
.
- β
-
33. A method for recursive filtering of time discrete measurement data relating to a measurable time-varying attribute to generate a reliability of the measurement data at a particular time instant, comprising the steps of:
-
collecting at least one measurement of the time-varying attribute during said time instant, wherein each said measurement has associated therewith a reliability indicator representing a relative reliability of said measurement;
filtering the reliability indicators for the at least one measurement to produce a filtered reliability indicator, said filtered reliability indicator including a time-weighted historical reliability indicator, wherein the historical reliability indicator includes information relating to a relative reliability of at least one measurement collected prior to said time instant. - View Dependent Claims (34)
-
-
35. An apparatus for recursively filtering measurement data, comprising:
-
a buffer for storing a plurality of measurements of a time-varying attribute, wherein each said measurement has associated therewith a reliability indicator representing a relative reliability of said measurement, each of said measurements of the time-varying attribute made during a particular time instant;
a memory for storing filtering instructions, said filtering instructions used to generate an estimated magnitude for the time-varying attribute at the particular time instant; and
a processor for filtering the plurality of measurements in accordance with the filtering instructions, said processor operating to;
filter the reliability indicators associated with the plurality of measurements to produce a filtered reliability indicator, said filtered reliability indicator determined using a time-weighted historical reliability indicator, wherein the historical reliability indicator includes information relating to a relative reliability of at least one measurement collected prior to said time instant;
time-weight a historical estimated magnitudes wherein the historical estimated magnitude includes information relating to at least one measurement of the time-varying attribute collected prior to the particular time instant;
reliability-weight the historical estimated magnitude using the filtered reliability indicator;
reliability-weight the plurality of measurements using the filtered reliability indicator to produce a reliability-weighted average of the plurality of measurements; and
generate an estimated magnitude of the time-varying attribute for the particular time instant using the reliability-weighted and time-weighted historical estimated magnitude and the reliability-weighted average of the plurality of measurements.
-
Specification