Granted Patent B2

US 12,560,944 · App. 17/138,285 · Granted Feb 24, 2026

Inventors: Thomas O. Teisberg (Menlo Park, CA); Rohit H. Sant (San Mateo, CA); Matthew O. Derry (Ann Arbor, MI); Michael J. Demertzi (Half Moon Bay, CA); Gavin K. Ananda Krishnan (San Carlos, CA); Keenan A. Wyrobek (Half Moon Bay, CA); Vasumathi Raman (San Francisco, CA); Brendan J.D. Wade (San Francisco, CA); Blair R. Hagen (Pacifica, CA); Randall R. Patterson (Littleton, CO); Philip M. Green (Sunnyvale, CA)

Assignee: SCIENTIFIC APPLICATIONS AND RESEARCH ASSOCIATES, INC.

G05D1/1064G05D1/0094G06N3/02H04R1/326H04R29/005H04R2499/13

Abstract

Claims (39)

1 . A method comprising:

receiving, at an acoustic detection system of an aircraft, audio signals, to provide a multichannel audio signal, wherein the acoustic detection system comprises two or more microphones, and wherein the audio signals are received by the acoustic detection system when the aircraft is traveling along a flight path based on an original flight plan:

determining, by one or more processors of the acoustic detection system, based on the multichannel audio signal, that the multichannel audio signal is associated with at least one intruding aircraft;

detecting, based on the multichannel audio signal, a second signal collected by a sensor of the aircraft, the second signal providing location information about the intruding aircraft;

based on the multichannel audio signal, the second signal, and the flight path, generating a plurality of possible maneuvers for the aircraft;

selecting a maneuver out of the plurality of possible maneuvers for execution by the aircraft to increase information detectable by the acoustic detection system, wherein the information is about the at least one intruding aircraft; and

commanding the maneuver for the aircraft based on the multichannel audio signal and the second signal, wherein the maneuver comprises a change in positioning of the aircraft relative to the intruding aircraft to change characteristics of the multichannel audio signal to obtain additional location information about the intruding aircraft, and wherein the maneuver further comprises a return to the flight path based on the original flight plan.

2 . The method of claim 1 , wherein selecting the maneuver of the plurality of possible maneuvers comprises:

selecting the maneuver for the aircraft using a machine learning model, wherein the machine learning model receives at least a track of the intruding aircraft as input, wherein the track is generated from two or more location estimations of the intruding aircraft.

3 . The method of claim 2 , wherein the machine learning model is a deep learning network generated using simulated flight data for the aircraft.

4 . The method of claim 2 , wherein the track of the intruding aircraft is generated based on an acoustic location estimation for the intruding aircraft, wherein the acoustic location estimation is generated using the multichannel audio signal.

5 . The method of claim 4 , wherein the acoustic location estimation is generated by the machine learning model using an input based on the multichannel audio signal.

6 . The method of claim 1 , wherein determining that the multichannel audio signal is associated with at least one intruding aircraft comprises determining that the multichannel audio signal is associated with the intruding aircraft and a second intruding aircraft, wherein commanding the maneuver is further based on a track of the second intruding aircraft generated based on the multichannel audio signal and additional information about the second intruding aircraft.

7 . The method of claim 1 , wherein the maneuver is selected by the acoustic detection system to increase or maintain a signal to noise ratio of the multichannel audio signal.

8 . The method of claim 1 , wherein generating the plurality of possible maneuvers for the aircraft comprises generating a maneuver configured to avoid the at least one intruding aircraft and gather information of the at least one intruding aircraft.

9 . The method of claim 1 , wherein selecting the plurality of possible maneuvers for the aircraft comprises selecting a maneuver based further on at least one of a weather data, flight condition, airspace restriction, and fuel level of the aircraft.

10 . The method of claim 1 , wherein selecting the maneuver out of the plurality of possible maneuvers for execution by the aircraft to increase information detectable by the acoustic detection system, wherein the information is about the at least one intruding aircraft, comprises:

assigning a score for each maneuver of the plurality of possible maneuvers, wherein a higher score indicates a higher amount of information about the at least one intruding aircraft detectable by the acoustic detection system of the at least one intruding aircraft during execution of the maneuver; and

selecting a maneuver of the plurality of possible maneuvers with the highest generated score.

11 . The method of claim 10 , wherein selecting the maneuver of the plurality of possible maneuvers further comprises:

evaluating whether each maneuver of the plurality of possible maneuvers will cause the aircraft to avoid the at least one intruding aircraft when executed by the aircraft; and

selecting a maneuver of the plurality of possible maneuvers that will cause the aircraft to avoid the at least one intruding aircraft.

12 . The method of claim 10 , wherein assigning the score for each maneuver of the plurality of possible maneuvers comprises assigning the score using a machine learning model, wherein the machine learning model is trained to classify the amount of information detectable by the acoustic detection system, wherein the information is about the at least one intruding aircraft, during execution of a maneuver.

13 . The method of claim 1 , wherein determining that the multichannel audio signal is associated with at least one intruding aircraft comprises:

processing the multichannel audio signal; and determining the location of the source of the multichannel audio signal using a localization machine learning model.

14 . The method of claim 13 , wherein processing the multichannel audio signal comprises transforming the multichannel audio signal to the frequency domain to obtain frequency data of the multichannel audio signal.

15 . The method of claim 13 , wherein the localization machine learning model is trained on labeled multichannel audio and directional data.

16 . The method of claim 1 , wherein determining that the multichannel audio signal is associated with at least one intruding aircraft comprises:

analyzing the multichannel audio signal to determine a class of the at least one intruding aircraft; and

determining a track of the at least one intruding aircraft based on the multichannel audio signal.

17 . The method of claim 16 , wherein analyzing the multichannel audio signal to determine a class of the at least one intruding aircraft comprises:

analyzing multichannel audio signal to determine a probability that the signal is associated with a class of aircraft; and

determining a flight characteristic associated with the class of aircraft.

18 . The method of claim 16 , wherein determining the track of the at least one intruding aircraft comprises:

determining a state estimation of the at least one intruding aircraft based on the multichannel audio signal; and

generating a track of the at least one intruding aircraft based on the state estimation of the at least one intruding aircraft and the class of the at least one intruding aircraft.

19 . The method of claim 16 , wherein determining the track of the at least one intruding aircraft comprises:

generating acoustic directional information corresponding to the at least one intruding aircraft using the multichannel audio signal; and

generating a track of the at least one intruding aircraft using the acoustic directional information and the second signal.

Assignments (2)

Continuity (5)

References Cited (217)