Adaptive cross-correlation
Techniques for adaptive cross-correlation are discussed. A first signal is received from a first audio sensor associated with a vehicle and a second signal is received from a second audio sensor associated with the vehicle. Techniques may include determining, based at least in part on the first signal, a first transformed signal in a frequency domain. Additionally, the techniques include determining, based at least in part on the second signal, a second transformed signal in the frequency domain. A parameter can be determined based at least in part on a characteristic associated with at least one of the vehicle, an environment proximate the vehicle, or one or more of the first or second signal. Cross-correlation data can be determined based at least in part on one or more of the first transformed signal, the second transformed signal, or the parameter.
1. A system comprising:
one or more processors; and
one or more non-transitory computer-readable media storing instructions executable by the one or more processors, wherein the instructions, when executed, cause the system to perform operations comprising:
receiving a first signal from a first audio sensor;
receiving a second signal from a second audio sensor;
determining, based at least in part on the first signal, a first transformed signal in a frequency domain;
determining, based at least in part on the second signal, a second transformed signal in the frequency domain;
determining, based at least in part on a characteristic associated with a vehicle and one or more of the first signal or the second signal, a parameter usable to modify a cross-correlation algorithm;
determining, based at least in part on the first transformed signal, the second transformed signal, and the parameter, cross-correlation data; and
determining, based at least in part on the cross-correlation data, a direction of arrival (DoA) value associated with the first signal and the second signal.
2. The system of claim 1 , wherein:
determining the parameter is further based on one or more of:
map data associated with a planned vehicle route,
a vehicle configuration,
environmental noise, or
a time of day.
3. The system of claim 1 , wherein determining the parameter comprises inputting one or more of the characteristic or information associated with an environment proximate the vehicle into a machine learned model and receiving, from the machine learned model, the parameter.
4. The system of claim 1 , wherein:
determining the parameter is further based on at least one of a signal-to-noise ratio, a sparseness, or harmonics associated with the first signal and the second signal.
5. The system of claim 1 , the operations further comprising:
dynamically determining an updated value of the parameter based on changing characteristics associated with the first signal or the second signal.
6. A method comprising:
receiving a first signal from a first audio sensor associated with a vehicle;
receiving a second signal from a second audio sensor associated with the vehicle;
determining, based at least in part on the first signal, a first transformed signal in a frequency domain;
determining, based at least in part on the second signal, a second transformed signal in the frequency domain;
determining, based at least in part on a characteristic associated with at least one of the vehicle or one or more of the first or the second signal, a parameter usable to modify a cross-correlation algorithm; and
determining, based at least in part on the first transformed signal, the second transformed signal, and the parameter, cross-correlation data.
7. The method of claim 6 , further comprising determining, based at least in part on the cross-correlation data, a direction of arrival (DoA) value associated with the first signal and the second signal.
8. The method of claim 6 , wherein determining the parameter is based at least in part on at least one of:
a frequency sparseness of at least one of the first transformed signal or the second transformed signal;
environmental noise proximate the vehicle;
a window size associated with at least one of the first signal or the second signal;
a fundamental frequency value associated with at least one of the first signal or the second signal; or
a signal-to-noise ratio of the first and second signals.
9. The method of claim 6 , further comprising:
inputting at least one of the first signal, the second signal, the first transformed signal, or the second transformed signal into a machine learned model; and
receiving, from the machine learned model, the parameter.
10. The method of claim 6 , further comprising:
initiating an activity based at least in part on a direction of arrival associated with the first signal and the second signal,
wherein the activity comprises at least one of contacting a remote computing system, changing a threshold in a perception algorithm, or initiating a vehicle maneuver.
11. The method of claim 6 , wherein determining the parameter is further based on at least one of map data associated with a planned vehicle route, a vehicle configuration, environmental noise, or a time of day.
12. The method of claim 6 , wherein determining the parameter is further based on at least one of a signal-to-noise ratio, a sparseness, or harmonics associated with at least one of the first signal or the second signal.
13. The method of claim 6 , further comprising dynamically determining an updated value of the parameter based on changing characteristics associated with the first signal or the second signal.
14. The method of claim 6 , further comprising determining a first parameter associated with the first audio sensor and determining a second parameter associated with the second audio sensor.
15. The method of claim 6 , further comprising:
applying multiple values of the parameter to the first signal and the second signal simultaneously; and
determining a specific parameter that produces a highest confidence level.
16. The method of claim 15 , wherein the specific parameter that produces the highest confidence level is used for processing the first signal and the second signal.
17. One or more non-transitory computer-readable media storing instructions that, when executed, cause one or more processors to perform operations comprising:
receiving a first signal from a first audio sensor associated with a vehicle;
receiving a second signal from a second audio sensor associated with the vehicle;
determining, based at least in part on the first signal, a first transformed signal in a frequency domain;
determining, based at least in part on the second signal, a second transformed signal in the frequency domain;
determining, based at least in part on a characteristic associated with at least one of the vehicle, an environment proximate the vehicle, or one or more of the first signal or the second signal, a parameter usable to modify a cross-correlation algorithm; and
determining, based at least in part on the first transformed signal, the second transformed signal, and the parameter, cross-correlation data.
18. The one or more non-transitory computer-readable media of claim 17 , wherein the operations further comprise:
determining, based at least in part on the cross-correlation data, a direction of arrival (DoA) value associated with the first signal and the second signal.
19. The one or more non-transitory computer-readable media of claim 17 , wherein determining the parameter is further based on at least one of map data associated with a planned vehicle route, a vehicle configuration, environmental noise, or a time of day.
20. The one or more non-transitory computer-readable media of claim 17 , wherein determining the parameter is further based on at least one of a signal-to-noise ratio, a sparseness, or harmonics associated with at least one of the first signal or the second signal.
21. The system of claim 1 , wherein:
the parameter is a cross-correlation adjusting parameter; and
the cross-correlation data is determined by a cross-correlation determination performed with the cross-correlation adjusting parameter.