초록 ▼

A method to initialize tracks from sensor measurements is provided. The method includes identifying at least one tentative track based on data collected from at least one sensor at three sequential times; initializing a confirm/delete track filter for the identified tentative tracks; and using gates

A method to initialize tracks from sensor measurements is provided. The method includes identifying at least one tentative track based on data collected from at least one sensor at three sequential times; initializing a confirm/delete track filter for the identified tentative tracks; and using gates computed from state vector statistics to one of: confirm the at least one tentative track; reprocess the at least one tentative track; or delete the at least one tentative track.

대표청구항 ▼

1. A method to initialize tracks from sensor measurements, the method comprising: identifying, at a processor, at least one tentative track by implementing a 3/3 measurement-to-measurement data association algorithm based on data collected from at least one sensor at at least three sequential times,

1. A method to initialize tracks from sensor measurements, the method comprising: identifying, at a processor, at least one tentative track by implementing a 3/3 measurement-to-measurement data association algorithm based on data collected from at least one sensor at at least three sequential times, wherein the implementing the 3/3 measurement-to-measurement data association algorithm comprises: resolving measurements from a spherical frame to a Cartesian frame;computing state vector statistics at an initial time;computing state vector statistics at a second time; andoutputting initial state vector statistics to initialize the confirm/delete track filter for the identified tentative track;identifying measurements at a first time that lie within a spherical gate of unassociated measurements from the initial time, wherein the spherical gate is a spherical region about a measurement vector and a gate threshold is a radius of the spherical region; andidentifying measurements at the second time that lie within the spherical gate of associated measurements from the first time,the method further comprising:initializing a confirm/delete track filter for the identified tentative tracks; andusing ellipsoidal gates computed from the state vector statistics to one of: confirm the at least one tentative track; reprocess the at least one tentative track; or delete the at least one tentative track. 2. The method of claim 1, wherein implementing the 3/3 measurement-to-measurement data association algorithm further comprises: identifying measurements at a first time that lie within a spherical gate of unassociated measurements from the initial time; andidentifying measurements at the second time that lie within the spherical gate of associated measurements from the first time. 3. The method of claim 1, wherein computing the state vector statistics at the initial time comprises performing a least squares analysis on a sequence of measurements and statistics associated with the sequence of measurements. 4. The method of claim 3, wherein computing the state vector statistics at the second time comprises propagating the initial state vector to the second time. 5. The method of claim 1, wherein using the ellipsoidal gates computed from the state vector statistics to confirm the at least one tentative track comprises: predicting state vector statistics at a third time by propagating the state vector statistics at a second time to the third time;implementing the ellipsoidal gates to determine if a candidate measurement is associated with a tentative track; andproviding a measurement assignment to the tentative track. 6. The method of claim 5, further comprising: updating confirmed tentative tracks;initializing and updating a score for the confirmed tentative track; andimplementing a threshold test on the updated confirmed tentative track and a track score. 7. The method of claim 1, wherein using the ellipsoidal gates computed from the state vector statistics to delete the at least one tentative track comprises: predicting state vector statistics at a third time by propagating the state vector statistics at a second time to the third time;implementing the ellipsoidal gates to determine if a candidate measurement is unassociated with a tentative track; andinitializing the confirm/delete track filter for the unassociated candidate measurement to determine if the unassociated candidate measurement is a new track. 8. A non-transitory program product comprising program instructions, embodied on a storage medium, that are operable to cause a processor to: identify at least one tentative track by implementing a 3/3 measurement-to-measurement data association algorithm based on data collected from at least one sensor at three sequential times, wherein the implementing the 3/3 measurement-to-measurement data association algorithm comprises program instructions, embodied on the storage medium, that are operable to cause the processor to: resolve measurements from a spherical frame to a Cartesian frame;compute state vector statistics at an initial time;compute state vector statistics at a second time; andoutput initial state vector statistics to initialize the confirm/delete track filter for the identified tentative track;identify measurements at a first time that lie within a spherical gate of unassociated measurements from the initial time, wherein the spherical gate is a spherical region about a measurement vector and a gate threshold is a radius of the spherical region; andidentify measurements at the second time that lie within the spherical gate of associated measurements from the first time;wherein the program instructions, embodied on the storage medium, are further operable to cause the processor to:initialize a confirm/delete track filter for the identified tentative tracks; anduse gates computed from the state vector statistics to one of confirm the at least one tentative track; reprocess the at least one tentative track; or delete the at least one tentative track. 9. The non-transitory program-product of claim 8, wherein the instructions operable to cause the processor to implement a 3/3 measurement-to-measurement data association algorithm further comprise instructions operable to cause the processor to: identify measurements at a first time that lie within a spherical gate of unassociated measurements from the initial time; andidentify measurements at a second time that lie within the spherical gate of associated measurements from the first time. 10. The non-transitory program-product of claim 8, wherein the instructions operable to cause the processor to compute the state vector statistics at the initial time comprise: perform a least squares analysis on a sequence of measurements and statistics associated with the sequence of measurements. 11. The non-transitory program-product of claim 10, wherein the instructions operable to cause the processor to compute the state vector statistics at the second time comprise instructions operable to cause the processor to propagate the initial state vector to the second time. 12. The non-transitory program-product of claim 8, wherein the instructions operable to cause the processor to use the gates computed from the state vector statistics to confirm the at least one tentative track comprise instructions operable to cause the processor to: predict state vector statistics at a third time by propagating the state vector statistics at a second time to the third time;implement ellipsoidal gates to determine if a candidate measurement is associated with a tentative track;initialize a track score for a confirmed tentative track; andprovide a measurement assignment to the tentative track, wherein the instructions further comprise instructions operable to cause the processor to:update confirmed tentative tracks and the track score; andimplement a threshold test on the updated confirmed tentative tracks and the track score. 13. The non-transitory program-product of claim 8, wherein the instructions operable to cause the processor to use the gates computed from the state vector statistics to delete the at least one tentative track comprise instructions operable to cause the processor to: predict state vector statistics at a third time by propagating the state vector statistics at a second time to the third time;implement ellipsoidal gates to determine if a candidate measurement is unassociated with a tentative track; andinitialize the confirm/delete track filter for the unassociated candidate measurement to determine if the unassociated candidate measurement is a new track. 14. A vehicle tracking system comprising: a processor communicatively coupled to input data from at least one sensor sensing objects in an environment of a vehicle; anda confirm/delete track filter including at least one algorithm operable to initialize a new track based on input from the at least one sensor, wherein the at least one algorithm operable to initialize the new track comprises:a first algorithm to identify at least one tentative track based on data collected from the at least one sensor at three sequential times by:resolving measurements from a spherical frame to a Cartesian frame;computing state vector statistics at an initial time;computing state vector statistics at a second time; andoutputting initial state vector statistics to initialize the confirm/delete track filter for the identified tentative track;identifying measurements at a first time that lie within a spherical gate of unassociated measurements from the initial time, wherein the spherical gate is a spherical region about a measurement vector and a gate threshold is a radius of the spherical region; andidentifying measurements at the second time that lie within the spherical gate of associated measurements from the first time;a second algorithm to initialize a confirm/delete track filter for the identified tentative tracks; anda third algorithm to use ellipsoidal gates computed from the state vector statistics to one of: confirm the at least one tentative track; reprocess the at least one tentative track; or delete the at least one tentative track. 15. The vehicle tracking system of claim 14, further comprising: the at least one sensor.