System and method for fusing asynchronous sensor tracks in a track fusion application
An example method can include receiving, at a sensor, a signal associated with a motion of a target, processing the signal via a first filter having a first motion model and a second filter having a second motion model to yield a first tracking output and a second tracking output for the target, and weighting the first tracking output and second tracking output according to how well each of the first motion model and second motion model represents the motion of the target, to yield a first weight for the first tracking output and a second weight for the second tracking output. The method can include combining the first tracking output and second tracking output to yield a fused tracking output and sending, to a fusion system, the fused tracking output, the first weight associated with the first tracking output and the second weight associated with the second tracking output.
1. A method comprising:
generating, via a first sensor, a first group of output tracks associated with a motion of a target object;
generating, via a second sensor, a second group of output tracks associated with the motion of the target object;
associating, via a track association module, a first group of similar tracks from the first group of output tracks and the second group of output tracks to yield a first set of associated output tracks;
associating, via the track association module, a second group of similar tracks from the first group of output tracks and the second group of output tracks to yield a second set of associated output tracks;
fusing, via a track fusion module, the first group of similar tracks into a first fused track; and
fusing, via the track fusion module, the second group of similar tracks into a second fused track.
2. The method of claim 1 , wherein:
the first sensor comprises a first filter having a first motion model and a second filter having a second motion model, to yield a first tracking output calculated for the target object using the first filter and a second tracking output calculated for the target object using the second filter, the first tracking output and the second tracking output being part of the first group of output tracks; and
the second sensor comprises a third filter having a third motion model and a fourth filter having a fourth motion model, to yield a third tracking output calculated for the target object using the third filter and a fourth tracking output calculated for the target object using the fourth filter, the third tracking output and the fourth tracking output being part of the second group of output tracks.
3. The method of claim 2 , further comprising:
weighting the first group of output tracks and the second group of output tracks, to yield a set of weights comprising a first weight associated with the first group of output tracks and a second weight associated with the second group of output tracks; and
initializing at least one filter using the fused tracking output and using at least one of the first fused track, the second fused track and the set of weights to perform an interactive multiple model propagation to a future time.
4. The method of claim 1 , further comprising:
aligning the first group of output tracks and the second group of output tracks in time to yield a first time aligned track and a second time aligned track, wherein the first group of output tracks and the second group of output tracks are aligned by projecting the first group of output tracks and the second group of output tracks to a common time step based on a first weight associated with the first group of output tracks and a second weight associated with the second group of output tracks; and
associating the first time aligned track with the second time aligned track based on a determination that the first time aligned track and the second time aligned track represent a same target object.
5. The method of claim 4 , further comprising:
after associating the first time aligned track and the second time aligned track, combining the first time aligned track and the second time aligned track to yield a single fused track associated with the target object.
6. The method of claim 1 , wherein a sum of all weights associated with the first fused track and the second fused track is equal to 1.
7. The method of claim 1 , wherein the first group of output tracks and the second group of output tracks comprise respective state and error covariance variables.
8. The method of claim 7 , wherein the respective state and error covariance variables comprise at least one of respective velocity variables, respective position variables, respective motion variables, and respective angle variables.
9. The method of claim 1 , wherein the first sensor is of a first type and the second sensor is of a second type.
10. The method of claim 9 , wherein the first type and the second type each relate to one or more of a motion sensor, a radio frequency sensor, a sound sensor, an optical sensor, a rangefinder, a radar sensor and an automatic dependent surveillance-broadcast sensor.
11. A system comprising:
one or more processors;
a first sensor;
a second sensor;
a track association module;
a track fusion module; and
at least one computer-readable storage medium having instructions stored which, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
generating, via the first sensor, a first group of output tracks associated with a motion of a target object;
generating, via the second sensor, a second group of output tracks associated with the motion of the target object;
associating, via the track association module, a first group of similar tracks from the first group of output tracks and the second group of output tracks to yield a first set of associated output tracks;
associating, via the track association module, a second group of similar tracks from the first group of output tracks and the second group of output tracks to yield a second set of associated output tracks;
fusing, via the track fusion module, the first group of similar tracks into a first fused track; and
fusing, via the track fusion module, the second group of similar tracks into a second fused track.
12. The system of claim 11 , wherein:
the first sensor comprises a first filter having a first motion model and a second filter having a second motion model, to yield a first tracking output calculated for the target object using the first filter and a second tracking output calculated for the target object using the second filter, the first tracking output and the second tracking output being part of the first group of output tracks; and
the second sensor comprises a third filter having a third motion model and a fourth filter having a fourth motion model, to yield a third tracking output calculated for the target object using the third filter and a fourth tracking output calculated for the target object using the fourth filter, the third tracking output and the fourth tracking output being part of the second group of output tracks.
13. The system of claim 12 , further comprising:
weighting the first group of output tracks and the second group of output tracks, to yield a set of weights comprising a first weight associated with the first group of output tracks and a second weight associated with the second group of output tracks; and
initializing at least one filter using the fused tracking output and using at least one of the first fused track, the second fused track and the set of weights to perform an interactive multiple model propagation to a future time.
14. The system of claim 11 , further comprising:
aligning the first group of output tracks and the second group of output tracks in time to yield a first time aligned track and a second time aligned track, wherein the first group of output tracks and the second group of output tracks are aligned by projecting the first group of output tracks and the second group of output tracks to a common time step based on a first weight associated with the first group of output tracks and a second weight associated with the second group of output tracks; and
associating the first time aligned track with the second time aligned track based on a determination that the first time aligned track and the second time aligned track represent a same target object.
15. The system of claim 14 , further comprising:
after associating the first time aligned track and the second time aligned track, combining the first time aligned track and the second time aligned track to yield a single fused track associated with the target object.
16. The system of claim 11 , wherein a sum of all weights associated with the first fused track and the second fused track is equal to 1.
17. The system of claim 11 , wherein the first group of output tracks and the second group of output tracks comprise respective state and error covariance variables.
18. The system of claim 17 , wherein the respective state and error covariance variables comprise at least one of respective velocity variables, respective position variables, respective motion variables, and respective angle variables.
19. The system of claim 11 , wherein the first sensor is of a first type and the second sensor is of a second type.
20. The system of claim 19 , wherein the first type and the second type each relate to one or more of a motion sensor, a radio frequency sensor, a sound sensor, an optical sensor, a rangefinder, a radar sensor and an automatic dependent surveillance-broadcast sensor.