US 2025/0035474 A1Published application

Methods and systems for filtering and estimating characteristics of an input signal using a tracker comprising a prism filter

AssigneesOxford University Innovation Limited, Coventry University

InventorManus Patrick HENRY

Life of this patent

Estimating the term…

Loading patent…

Abstract

A method of estimating characteristics of an input signal comprises inputting the input signal into a first tracker to determine a first estimate of characteristics of the input signal, and determining a normalised signal from the input signal. The normalised signal is input into a second tracker to determine a second estimate of characteristics of the input signal. At least one of the first tracker and the second tracker comprises or is derived from at least one Prism filter. Further, a method of filtering an input signal is provided. The method comprises applying a sequence of one or more filter stages to the input signal to generate respective downsampled signals. A tracker stage is applied to the output of the final filter stage to determine one or more characteristics of the input signal.

Claims25

Dependent claims

A method of estimating characteristics of an input signal, the method comprising:

inputting the input signal into a first tracker to determine a first estimate of characteristics of the input signal, the first estimate comprising a first amplitude estimate and a first frequency estimate,
determining a normalised signal from the input signal by:
normalising the amplitude of the input signal using the first amplitude estimate;
and/or
heterodyning the input signal using a heterodyne frequency determined based on a sum and/or difference between the first frequency estimate and a filter frequency;
and
inputting the normalised signal into a second tracker to determine a second estimate of characteristics of the input signal, the second tracker comprising at least one Prism filter, wherein the second tracker is configured to pass a narrower range of frequencies than the first tracker, and wherein the second tracker is configured to pass the filter frequency;
wherein at least one of the first tracker and the second tracker comprises or is derived from at least one Prism filter, wherein:
each Prism filter is configured to evaluate double integrals of the form:
I
[
S
|
C
]
[
S
|
C
]
h
=
m
2
⁢
∫
-
1
m
0
[
sin
|
cos
]
⁢
(
2
⁢
π
⁢
hmt
+
q
)
⁢
(
∫
t
-
1
m
t
[
sin
|
cos
]
⁢
(
2
⁢
π
⁢
hmt
+
q
)
·
s
⁡
(
t
)
⁢
dt
)
⁢
dt
,
where s(t) is an input to the respective Prism filter, m is a characteristic frequency of the respective Prism filter, [sin|cos] represents a sine or a cosine operation, and q is an optional phase offset variable, and h is a harmonic number of the respective Prism filter.

78 more elements

A method of filtering an input signal, the method comprising:

applying a sequence of one or more filter stages to the input signal, wherein:
the input to a first filter stage in the sequence is the input signal, and the input to any subsequent filter stages in the sequence is a respective downsampled signal generated by an immediately preceding filter stage;
and
applying a respective filter stage of the sequence of one or more filter stages comprises:
heterodyning the input for the respective filter stage based on a respective target frequency and a respective filter frequency;
applying a respective convolutional filter to a subset of samples of the heterodyned input selected based on a respective downsampling factor, the respective convolutional filter configured to pass the respective filter frequency;
generating a respective downsampled signal from outputs of the respective convolutional filter applied to the subset of samples;
and
applying a tracker stage to the downsampled signal generated by a final filter stage in the sequence of one or more filter stages, wherein applying the tracker stage comprises:
heterodyning the downsampled signal generated by the final filter stage based on a tracker target frequency and a tracker filter frequency;
applying a tracker filter to the heterodyned downsampled signal generated by the final filter stage to determine one or more characteristics of the input signal;
wherein each respective convolutional filter is derived from a network of Prism filters comprising at least one Prism filter, and the tracker filter comprises or is derived from at least one Prism filter, wherein each Prism filter is configured to evaluate double integrals of the form:
I
[
S
|
C
]
[
S
|
C
]
h
=
m
2
⁢
∫
-
1
m
0
[
sin
|
cos
]
⁢
(
2
⁢
π
⁢
hmt
+
q
)
⁢
(
∫
t
-
1
m
t
[
sin
|
cos
]
⁢
(
2
⁢
π
⁢
hmt
+
q
)
·
s
⁡
(
t
)
⁢
dt
)
⁢
dt
,
where s(t) is an input to the respective Prism filter, m is a characteristic frequency of the respective Prism filter, [sin|cos] represents a sine or a cosine operation, and q is an optional phase offset variable, and h is a harmonic number of the respective Prism filter.

81 more elements

Dependencies are inferred from the claim text.