Compact machine learning architecture for wideband direction finding with circular antenna arrays

1 . A computer system for angle of arrival estimation comprising:

one or more processors; and

one or more computer-readable media having stored thereon executable instructions that when executed by the one or more processors configure the computer system to:

receive a single snapshot from a circular array of antennas, wherein the antennas are configured to receive a signal and circularly arranged in a circle;

process the single snapshot for finding a direction of the signal using two cascaded algorithms; and

perform, by a first algorithm of the two cascaded algorithms, a circular data-shift operation, which corresponds to a circular configuration of the circular array of antennas, only on an amplitude vector of the single snapshot to identify a target subregion, from which a signal arrives, where the target subregion is defined by an azimuthal angle and separated from neighboring subregions; and

identify, by a second algorithm of the two cascaded algorithm, an angle of arrival of the signal within the subregion, wherein:

the angle of arrival is independent of the azimuthal angle of the target subregion.

2 . The computer system as recited in claim 1 , wherein the first algorithm comprises a first machine learning algorithm, wherein the first machine learning algorithm is trained to identify the target subregion from which the signal arrives.

3 . The computer system as recited in claim 1 , wherein the second algorithm comprises a second machine learning algorithm, wherein the second machine learning algorithm is trained to identify the angle of arrival of the signal within the target subregion.

4 . The computer system as recited in claim 1 , wherein the second algorithm identifies the angle of arrival of the signal within any subregion received from the first algorithm.

5 . The computer system as recited in claim 1 , wherein the amplitude of the signal is an amplitude voltage or amplitude power readings.

6 . A computer-implemented method, executed on one or more processors, for angle of arrival estimation, the computer-implemented method comprising:

receiving a single snapshot from a circular array of antennas, wherein the antennas are configured to receive a signal and circularly arranged in a circle;

processing the single snapshot for finding a direction of the signal using two cascaded algorithms;

performing, by a first algorithm of the two cascaded algorithms, a circular data-shift operation, which corresponds to a circular configuration of the circular array of antennas, only on an amplitude vector of the single snapshot to identify a target subregion, from which a signal arrives, where the target subregion is defined by an azimuthal angle and separated from neighboring subregions; and

identifying, by a second algorithm of the two cascaded algorithm, an angle of arrival of the signal within the subregion, wherein:

the angle of arrival is independent of the azimuthal angle of the target subregion.

7 . The computer-implemented method as recited in claim 6 , wherein the first algorithm comprises a first machine learning algorithm, wherein the first machine learning algorithm is trained to identify the target subregion from which the signal arrives.

8 . The computer-implemented method as recited in claim 6 , wherein the second algorithm comprises a second machine learning algorithm, wherein the second machine learning algorithm is trained to identify the angle of arrival of the signal within the target subregion.

9 . The computer-implemented method as recited in claim 6 , wherein the second algorithm identifies the angle of arrival of the signal within any subregion received from the first algorithm.

10 . The computer-implemented method as recited in claim 6 , wherein the amplitude of the signal is an amplitude voltage or amplitude power readings.

11 . A computer-readable medium comprising one or more physical computer-readable storage media having stored thereon computer-executable instructions that, when executed at a processor, cause a computer system to perform a method for angle of arrival estimation, the method comprising:

receiving a single snapshot from a circular array of antennas, wherein the antennas are configured to receive a signal and circularly arranged in a circle;

processing the single snapshot for finding a direction of the signal using two cascaded algorithms;

performing, by a first algorithm of the two cascaded algorithms, a circular data-shift operation, which corresponds to a circular configuration of the circular array of antennas, only on an amplitude vector of the single snapshot to identify a target subregion, from which a signal arrives, where the target subregion is defined by an azimuthal angle and separated from neighboring subregions; and

identifying, by a second algorithm of the two cascaded algorithm, an angle of arrival of the signal within the subregion, wherein:

the angle of arrival is independent of the azimuthal angle of the target subregion.

12 . The computer-readable medium as recited in claim 11 , wherein the first algorithm comprises a first machine learning algorithm, wherein the first machine learning algorithm is trained to identify the target subregion from which the signal arrives.

13 . The computer-readable medium as recited in claim 11 , wherein the second algorithm comprises a second machine learning algorithm, wherein the second machine learning algorithm is trained to identify the angle of arrival of the signal within the target subregion.

14 . The computer-readable medium as recited in claim 11 , wherein the second algorithm identifies the angle of arrival of the signal within any subregion received from the first algorithm.