Environmental perception in autonomous driving using captured audio
This application discloses sensors to capture audio measurement in an environment around a vehicle and a computing system to classify audio measurements captured with one or more sensors mounted to a vehicle, wherein the classified audio measurements identify to a type of object in an environment around the vehicle, and fuse the classified audio measurements with measurements captured by at least one different type of sensor to detect the object in the environment around the vehicle, wherein a control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the detected object. The computing system can also identify noise in the captured audio measurements originating from the vehicle and utilize the identified noise to detect faults in the vehicle, to perform proximity detection around the vehicle, or to perform noise cancelation operations in the vehicle.
1. A method comprising:
capturing audio measurements with one or more sensors mounted to a vehicle;
classifying, by a computing system, the captured audio measurements as corresponding to a type of object in an environment around the vehicle; and
fusing, by the computing system, the classified audio measurements with measurements captured by at least one different type of sensor to detect the object in the environment around the vehicle, wherein a control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the detected object, and wherein, when the type of object associated with the classified captured audio measurements is not visible to the at least one different type of the sensor, the control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the type of object associated with the classified captured audio measurements.
2. The method of claim 1 , further comprising:
identifying, by the computing system, noise in the captured audio measurements originating from the vehicle; and
filtering, by the computing system, the captured audio measurement to remove the noise, wherein the classifying utilizes the captured audio measurements having the noise removed.
3. The method of claim 1 , further comprising:
identifying, by the computing system, noise in the captured audio measurements originating from the vehicle; and
utilizing, by the computing system, the noise in the captured audio measurement detect a fault in the vehicle.
4. The method of claim 1 , further comprising utilizing, by the computing system, the captured audio measurements to detect external weather conditions, wherein the control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the external weather conditions.
5. The method of claim 1 , further comprising:
utilizing, by the computing system, the captured audio measurements to detect a proximity of the vehicle to the object in the environment; and
generating, by the computing system, occupancy data to identify areas in the environment around the vehicle corresponding to free space based, at least in part, on the detected proximity of the vehicle to the object in the environment.
6. The method of claim 1 , further comprising utilizing, by the computing system, the captured audio measurements to detect a proximity of the vehicle to a ground in the environment, wherein the fusing of the classified audio measurements with measurements captured by at least one different type of sensor is performed relative to the detected proximity of the vehicle to the ground in the environment.
7. An apparatus comprising at least one memory device storing instructions configured to cause one or more processing devices to perform operations comprising:
classifying audio measurements captured with one or more sensors mounted to a vehicle, wherein the classified audio measurements identify to a type of object in an environment around the vehicle;
identifying noise in the captured audio measurements originating from the vehicle;
filtering the captured audio measurement to remove the noise, wherein the classifying utilizes the captured audio measurements having the noise removed; and
fusing the classified audio measurements with measurements captured by at least one different type of sensor to detect the object in the environment around the vehicle, wherein a control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the detected object.
8. The apparatus of claim 7 , wherein, when the type of object associated with the classified captured audio measurements is not visible to the at least one different type of the sensor, the control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the type of object associated with the classified captured audio measurements.
9. The apparatus of claim 7 , wherein the instructions are further configured to cause the one or more processing devices to perform operations comprising:
identifying noise in the captured audio measurements originating from the vehicle; and
utilizing the noise in the captured audio measurement detect a fault in the vehicle.
10. The apparatus of claim 7 , wherein the instructions are further configured to cause the one or more processing devices to perform operations comprising utilizing the captured audio measurements to detect external weather conditions, wherein the control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the external weather conditions.
11. The apparatus of claim 7 , wherein the instructions are further configured to cause the one or more processing devices to perform operations comprising:
utilizing the captured audio measurements to detect a proximity of the vehicle to the object in the environment; and
generating occupancy data to identify areas in the environment around the vehicle corresponding to free space based, at least in part, on the detected proximity of the vehicle to the object in the environment.
12. The apparatus of claim 7 , wherein the instructions are further configured to cause the one or more processing devices to perform operations comprising utilizing the captured audio measurements to detect a proximity of the vehicle to a ground in the environment, wherein the fusing of the classified audio measurements with measurements captured by at least one different type of sensor is performed relative to the detected proximity of the vehicle to the ground in the environment.
13. A system comprising:
a memory device configured to store machine-readable instructions; and
a computing system including one or more processing devices, in response to executing the machine-readable instructions, configured to:
classify audio measurements captured with one or more sensors mounted to a vehicle, wherein the classified audio measurements identify to a type of object in an environment around the vehicle;
utilize the captured audio measurements to detect a proximity of the vehicle to a ground in the environment; and
fuse the classified audio measurements with measurements captured by at least one different type of sensor relative to the detected proximity of the vehicle to the ground in the environment to detect the object in the environment around the vehicle, wherein a control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the detected object.
14. The system of claim 13 , wherein, when the type of object associated with the classified captured audio measurements is not visible to the at least one different type of the sensor, the control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the type of object associated with the classified captured audio measurements.
15. The system of claim 13 , wherein the one or more processing devices, in response to executing the machine-readable instructions, are configured to:
identify noise in the captured audio measurements originating from the vehicle; and
utilize the noise in the captured audio measurement detect a fault in the vehicle.
16. The system of claim 13 , wherein the one or more processing devices, in response to executing the machine-readable instructions, are configured to utilize the captured audio measurements to detect external weather conditions, wherein the control system for the vehicle is configured to control operation of the vehicle based, at least in part, on the external weather conditions.
17. The system of claim 13 , wherein the one or more processing devices, in response to executing the machine-readable instructions, are configured to:
utilize the captured audio measurements to detect a proximity of the vehicle to the object in the environment; and
generate occupancy data to identify areas in the environment around the vehicle corresponding to free space based, at least in part, on the detected proximity of the vehicle to the object in the environment.