Systems and methods for bayesian likelihood estimation of fused objects
A sensor fusion system and method are disclosed. One or more processors are operable to receive a plurality of object detection measurements from a plurality of sensors. Each of the plurality of object detection measurements are associated with a potential object detection track. A plurality of sensor confidence values associated with each of the plurality of sensors are received. A track confidence value is determined for each of the potential object detection tracks based on the received plurality of object detection measurements and the received plurality of sensor confidence values. An object detection for a potential object detection track that has a determined track confidence value meeting a predetermined detection threshold is then determined, or confirmed, and stored in a memory for subsequent use, and is relatively unaffected by a measurement from a sensor that has a field of view that omits or is occluded with respect to the given object detection track.
1. A sensor fusion method, comprising:
receiving a plurality of object detection measurements from a plurality of sensors, each of the plurality of object detection measurements being associated with a respective one of a plurality of potential object detection tracks, each of the plurality of sensors having a different field of view based on a corresponding location on a vehicle;
receiving a plurality of sensor confidence values associated with the plurality of sensors;
determining a track confidence value for each of the plurality of potential object detection tracks based on the plurality of object detection measurements and the plurality of sensor confidence values; and
fusing at least two of the plurality of potential object detection tracks corresponding to different fields of view into a true object detection for controlling at least a portion of the vehicle based at least in part on the track confidence value of each of the at least two of the plurality of potential object detection tracks meeting a predetermined detection threshold.
2. The sensor fusion method of claim 1 , further comprising one or more of:
discarding a potential object detection track of the plurality of potential object detection tracks that has a track confidence value meeting a predetermined non-detection threshold; and
receiving further object detection measurements from the plurality of sensors for a potential object detection track that has a determined track confidence value not meeting the predetermined detection threshold.
3. The sensor fusion method of claim 1 , wherein the track confidence value for a given potential object detection track is unaffected by a measurement from a sensor of the plurality of sensors that has a field of view that omits or is occluded with respect to the given potential object detection track.
4. The sensor fusion method of claim 1 , wherein determining the track confidence value for each of the plurality of potential object detection tracks based on the plurality of object detection measurements and the plurality of sensor confidence values comprises applying a Bayesian filtering algorithm to each of the plurality of potential object detection tracks, wherein the Bayesian filtering algorithm comprises instructions stored in a memory and executed by a processor.
5. The sensor fusion method of claim 1 , wherein each of the plurality of sensor confidence values associated with the plurality of sensors comprises one or more of a false positive rate, a false negative rate, a true positive rate, and a true negative rate.
6. The sensor fusion method of claim 1 , wherein each of the plurality of sensors comprises one of a camera, a radar sensor, a lidar sensor, and a vehicle state sensor.
7. The sensor fusion method of claim 1 , further comprising actuating one or more of an alert system, an acceleration system, a braking system, a steering system, and a suspension system of the vehicle based on the true object detection.
8. The sensor fusion method of claim 1 , wherein each of the plurality of sensor confidence values indicates a rate of detection between false positive and false negative for a respective one of the plurality of sensors.
9. A non-transitory computer-readable medium comprising instructions stored in a memory and executed by a processor to carry out steps of a sensor fusion method, comprising:
receiving a plurality of object detection measurements from a plurality of sensors, each of the plurality of sensors having a different field of view based on a corresponding location on a vehicle;
associating each of the plurality of object detection measurements with a respective one of a plurality of potential object detection tracks;
receiving a plurality of sensor confidence values associated with the plurality of sensors;
determining a track confidence value for each of the plurality of potential object detection tracks based on the plurality of object detection measurements and the plurality of sensor confidence values; and
fusing at least two of the plurality of potential object detection tracks corresponding to different fields of view into a true object detection for controlling at least a portion of the vehicle based at least in part on the track confidence value of each of the at least two of the plurality of potential object detection tracks meeting a predetermined detection threshold.
10. The non-transitory computer readable medium of claim 9 , wherein the steps further comprise one or more of:
discarding a potential object detection track of the plurality of potential object detection tracks that has a track confidence value meeting a predetermined non-detection threshold; and
receiving further object detection measurements from the plurality of sensors for a potential object detection track that has a determined track confidence value not meeting the predetermined detection threshold.
11. The non-transitory computer readable medium of claim 9 , wherein the determined track confidence value for a given potential object detection track is unaffected by a measurement from a sensor of the plurality of sensors that has a field of view that omits or is occluded with respect to the given potential object detection track.
12. The non-transitory computer readable medium of claim 9 , wherein determining the track confidence value for each of the plurality of potential object detection tracks based on the plurality of object detection measurements and the plurality of sensor confidence values comprises applying a Bayesian filtering algorithm to each of the plurality of potential object detection tracks.
13. The non-transitory computer readable medium of claim 9 , wherein each of the plurality of sensor confidence values associated with the plurality of sensors comprises one or more of a false positive rate, a false negative rate, a true positive rate, and a true negative rate.
14. The non-transitory computer readable medium of claim 9 , wherein each of the plurality of sensors comprises one of a camera, a radar sensor, a lidar sensor, and a vehicle state sensor.
15. The non-transitory computer readable medium of claim 9 , wherein the steps further comprise actuating one or more of an alert system, an acceleration system, a braking system, a steering system, and a suspension system of the vehicle based on the true object detection.
16. A sensor fusion system, comprising:
a plurality of sensors providing a plurality of object detection measurements, each of the plurality of sensors having a different field of view based on a corresponding location on a vehicle;
a processing unit; and
a memory storing instructions, which when executed by the processing unit, cause the processing unit to perform operations comprising:
associating each of the plurality of object detection measurements with a respective one of a plurality of potential object detection tracks;
receiving a plurality of sensor confidence values associated with the plurality of sensors;
determining a track confidence value for each of the plurality of potential object detection tracks based on the plurality of object detection measurements and the plurality of sensor confidence values; and
a fusing at least two of the plurality of potential object detection tracks corresponding to different fields of view into a true object detection for controlling at least a portion of the vehicle based at least in part on the track confidence value of each of the at least two of the plurality of potential object detection tracks meeting a predetermined detection threshold.
17. The sensor fusion system of claim 16 , wherein the operations further comprise:
discarding a potential object detection track of the plurality of potential object detection tracks that has a track confidence value meeting a predetermined non-detection threshold; and
receiving further object detection measurements from the plurality of sensors for a potential object detection track that has a determined track confidence value not meeting the predetermined detection threshold.
18. The sensor fusion system of claim 16 , wherein the track confidence value for a given potential object detection track is unaffected by a measurement from a sensor of the plurality of sensors that has a field of view that omits or is occluded with respect to the given potential object detection track.
19. The sensor fusion system of claim 16 , wherein determining the track confidence value for each of the plurality of potential object detection tracks based on the plurality of object detection measurements and the plurality of sensor confidence values comprises applying a Bayesian filtering algorithm to each of the plurality of potential object detection tracks.
20. The sensor fusion system of claim 16 , wherein each of the plurality of sensor confidence values associated with the plurality of sensors comprises one or more of a false positive rate, a false negative rate, a true positive rate, and a true negative rate.
21. The sensor fusion system of claim 16 , wherein each of the plurality of sensors comprises one of a camera, a radar sensor, a lidar sensor, and a vehicle state sensor.