Long-term object tracking supporting autonomous vehicle navigation

1. A method, comprising:

capturing sensor data at a first time using a sensor of an autonomous vehicle;

detecting, using a processing circuit, a first object from the sensor data, wherein detecting the first object includes creating detection data;

generating, using the processing circuit, first tracking data associated with the first object based on the detection data;

obtaining, using the processing circuit, second tracking data that meets a stale-track criteria, wherein the second tracking data corresponds to a second object detected using second data captured at a second time, wherein the second time is prior to the first time;

determining a predicted position of the second object at the first time based on the second tracking data;

determining that the first tracking data and the second tracking data satisfy matching criteria, wherein determining that the first tracking data and the second tracking data satisfy matching criteria comprises determining that the predicted position of the second object at the first time corresponds to a position of the first object at the first time; and

in accordance with determining that the first tracking data and the second tracking data satisfy the matching criteria:

associating, using the processing circuit, the second tracking data with the first tracking data; and

navigating, using a control circuit, the autonomous vehicle based at least in part on the second tracking data.

2. The method of claim 1 , wherein the detection data includes at least one of a detected range, range rate, angular position, or velocity.

3. The method of claim 1 , wherein the first tracking data is assigned a first track ID.

4. The method of claim 3 , wherein associating, using the processing circuit, the second tracking data with the first tracking data includes: assigning the first track ID to the second tracking data.

5. The method of claim 1 , further comprising predicting a future position of the object based at least in part on the first tracking data and the second tracking data.

6. The method of claim 5 , further comprising:

capturing additional sensor data;

generating new tracking data from the additional sensor data; and

in accordance with a determination that the new tracking data includes a position of the

object that corresponds to the predicted future position of the object, assigning a same track ID to the first tracking data and the new tracking data.

7. The method of claim 1 , wherein the stale-track criteria is based at least in part on tracking data age.

8. The method of claim 7 , wherein tracking data confidence of the second tracking data is based at least in part on a temporal consistency of the data making up the second tracking data.

9. The method of claim 1 , wherein the first tracking data further comprises at least one of a velocity of the object, a shape of the object, a size of the object, a color of the object, or an orientation of the object.

10. The method of claim 1 , wherein the second tracking data meets the stale-track criteria due to the occurrence of at least one of object occlusion, illumination variation, or deformation.

11. The method of claim, 1 , wherein the sensor data includes imagery and creating the detection data includes extracting visual features from portions of the imagery that include the object.

12. The method of claim 1 , wherein the sensor is a fixed focus imaging sensor.

13. The method of claim 1 , wherein the sensor is one of a LiDAR sensor and a RADAR sensor.

14. The method of claim 1 , wherein the tracking data is generated by at least one of a correlation filter tracker, a deep tracker, or a Kalman filter tracker.

15. The method of claim 1 , wherein the stale-track criteria is met when the most recent data associated with the tracking data is older than a first threshold period of time.

16. The method of claim 15 , further comprising:

in accordance with a determination that the most recent data associated with the tracking data is older than a second threshold period of time that is greater than the first threshold period of time, deleting tracking data categorized as being stale.

17. The method of claim 1 , wherein the comparison meets the matching criteria when at least one of an object shape, size, or position included in the first tracking data matches a shape, size, or position included in the second tracking data.

18. The method of claim 17 , wherein:

in accordance with a determination that differences in the first tracking data and the second tracking data are due to a change in orientation of the object with respect to the sensor, comparing the first tracking data to the second tracking data disregards the differences in the first tracking data and the second tracking data that are due to the change in orientation of the object with respect to the sensor.

19. A non-transitory computer-readable storage medium storing instructions configured to be executed by one or more circuits of a computing device that cause the computing device to carry out steps that include:

capturing sensor data at a first time using a sensor of an autonomous vehicle;

detecting, using a processing circuit, a first object from the sensor data, wherein detecting the first object includes creating detection data;

generating, using the processing circuit, first tracking data associated with the first object based on the detection data;

obtaining, using the processing circuit, second tracking data that meets a stale-track criteria, wherein the second tracking data corresponds to a second object detected using second data captured at a second time, wherein the second time is prior to the first time;

determining a predicted position of the second object at the first time based on the second tracking data;

determining that the first tracking data and the second tracking data satisfy matching criteria, wherein determining that the first tracking data and the second tracking data satisfy matching criteria comprises determining that the predicted position of the second object at the first time corresponds to a position of the first object at the first time; and

in accordance with determining that the first tracking data and the second tracking data satisfy the matching criteria:

associating, using the processing circuit, the second tracking data with the first tracking data; and

navigating, using a control circuit, the autonomous vehicle based at least in part on the second tracking data.

20. An autonomous vehicle, comprising:

a sensor;

a processing circuit;

a control circuit; and

memory storing one or more programs configured to be executed by the circuits of the autonomous vehicle, the one or more programs including instructions for:

capturing sensor data at a first time using the sensor;

detecting, using the processing circuit, a first object from the sensor data, wherein detecting the first object includes creating detection data;

generating, using the processing circuit, first tracking data associated with the first object based on the detection data;

obtaining, using the processing circuit, second tracking data that meets a staletrack criteria, wherein the second tracking data corresponds to a second object detected using second data captured at a second time, wherein the second time is prior to the first time;

determining a predicted position of the second object at the first time based on the second tracking data;

determining that the first tracking data and the second tracking data satisfy matching criteria, wherein determining that the first tracking data and the second tracking data satisfy matching criteria comprises determining that the predicted position of the second object at the first time corresponds to a position of the first object at the first time; and

in accordance with determining that the first tracking data and the second tracking data satisfy the matching criteria:

associating, using the processing circuit, the second tracking data with the first tracking data; and

navigating, using the control circuit, the autonomous vehicle based at least in part on the second tracking data.