Systems and methods for prediction of a jaywalker trajectory through an intersection

1. A method for generating a trajectory of an uncertain road user (URU) in proximity of a drivable area that a vehicle will traverse, the method comprising:

detecting a URU within a threshold distance of a drivable area that a vehicle is traversing or will traverse;

receiving perception information relating to the URU;

for each of a plurality of entry points on a drivable area boundary that the URU can use to enter the drivable area, using a plurality of features associated with that entry point to determine a likelihood that the URU will enter the drivable area from that entry point, the plurality of features computed based on the perception information;

generating, using the plurality of entry points and the corresponding likelihoods, a predicted trajectory of the URU; and

controlling, based on the predicted trajectory, navigation of the vehicle while traversing the drivable area to avoid collision with the URU,

wherein the plurality of features associated with that entry point comprise at least one of the following:

a distance between the URU's current position and that entry point wherein the likelihood that the URU will enter the drivable area from that entry point increases as the distance decreases,

time the URU will take to reach that entry point wherein the likelihood that the URU will enter the drivable area from that entry point increases as the time decreases,

a difference between a tangential vector of the drivable area boundary at that entry point and a heading of the URU wherein the likelihood that the URU will enter the drivable area from that entry point increases as the difference increases,

a distance between the URU's current position and a lane within the drivable area that is closest to that entry point wherein the likelihood that the URU will enter the drivable area from that entry point increases as the distance decreases, or

a difference between a direction of the lane and the heading of the URU wherein the likelihood that the URU will enter the drivable area from that entry point increases as the difference increases.

2. The method of claim 1 , further comprising identifying each of the plurality of entry points on the drivable area boundary that the URU can use to enter the drivable area by:

defining a plurality of hypotheses relating to the URU's trajectory for crossing the drivable area boundary; and

using the perception information and the plurality of hypotheses for identifying the plurality of entry points.

3. The method of claim 2 , wherein the plurality of hypotheses comprise at least one of the following:

the URU will enter the drivable area from an entry point on the drivable area boundary that is closest to the URU's current position;

the URU will enter the drivable area from an entry point on the drivable area boundary where the drivable area boundary intersects with a heading vector of the URU; or

the URU will enter the drivable area from an entry point that is determined based on historical information about the drivable area.

4. The method of claim 1 , further comprising predicting, using the predicted trajectory, whether the URU will enter the drivable area.

5. A system for generating a trajectory of an uncertain road user (URU) in proximity of a drivable area that a vehicle will traverse, the system comprising:

a processor and a memory containing programming instructions that are configured to cause the vehicle processor to:

detect a URU within a threshold distance of a drivable area that a vehicle is traversing or will traverse,

receive perception information relating to the URU,

for each of a plurality of entry points on a drivable area boundary that the URU can use to enter the drivable area, use a plurality of features associated with that entry point to determine a likelihood that the URU will enter the drivable area from that entry point, the plurality of features computed based on the perception information,

generate, using the plurality of entry points and the corresponding likelihoods, a predicted trajectory of the URU, and

control, based on the predicted trajectory, navigation of the vehicle while traversing the drivable area to avoid collision with the URU,

wherein the plurality of features associated with that entry point comprise at least one of the following:

a distance between the URU's current position and that entry point wherein the likelihood that the URU will enter the drivable area from that entry point increases as the distance decreases,

time the URU will take to reach that entry point wherein the likelihood that the URU will enter the drivable area from that entry point increases as the time decreases,

a difference between a tangential vector of the drivable area boundary at that entry point and a heading of the URU wherein the likelihood that the URU will enter the drivable area from that entry point increases as the difference increases,

a distance between the URU's current position and a lane within the drivable area that is closest to that entry point wherein the likelihood that the URU will enter the drivable area from that entry point increases as the distance decreases, or

a difference between a direction of the lane and the heading of the URU wherein the likelihood that the URU will enter the drivable area from that entry point increases as the difference increases.

6. The system of claim 5 , further comprising instructions that cause the processor to identify each of the plurality of entry points on the drivable area boundary that the URU can use to enter the drivable area by:

defining a plurality of hypotheses relating to the URU's trajectory for crossing the drivable area boundary; and

using the perception information and the plurality of hypotheses for identifying the plurality of entry points.

7. The system of claim 6 , wherein the plurality of hypotheses comprise at least one of the following:

the URU will enter the drivable area from an entry point on the drivable area boundary that is closest to the URU's current position;

the URU will enter the drivable area from an entry point on the drivable area boundary where the drivable area boundary intersects with a heading vector of the URU; or

the URU will enter the drivable area from an entry point that is determined based on historical information about the drivable area.

8. The system of claim 5 , further comprising further comprising instructions that cause the processor to predict, using the predicted trajectory, whether the URU will enter the drivable area.

9. A computer program product comprising a non-transitory computer readable medium that contains programming instructions that are configured to cause an autonomous vehicle to:

detect a URU within a threshold distance of a drivable area that a vehicle is traversing or will traverse;

receive perception information relating to the URU;

for each of a plurality of entry points on a drivable area boundary that the URU can use to enter the drivable area, use a plurality of features associated with that entry point to determine a likelihood that the URU will enter the drivable area from that entry point, the plurality of features computed based on the perception information;

generate, using the plurality of entry points and the corresponding likelihoods, a predicted trajectory of the URU; and

control, based on the predicted trajectory, navigation of the vehicle while traversing the drivable area to avoid collision with the URU,

wherein the plurality of features associated with that entry point comprise at least one of the following:

a distance between the URU's current position and that entry point wherein the likelihood that the URU will enter the drivable area from that entry point increases as the distance decreases,

time the URU will take to reach that entry point wherein the likelihood that the URU will enter the drivable area from that entry point increases as the time decreases,

a difference between a tangential vector of the drivable area boundary at that entry point and a heading of the URU wherein the likelihood that the URU will enter the drivable area from that entry point increases as the difference increases,

a distance between the URU's current position and a lane within the drivable area that is closest to that entry point wherein the likelihood that the URU will enter the drivable area from that entry point increases as the distance decreases, or

a difference between a direction of the lane and the heading of the URU wherein the likelihood that the URU will enter the drivable area from that entry point increases as the difference increases.

10. The computer program product of claim 9 , further comprising programming instructions that cause the autonomous vehicle to identify each of the plurality of entry points on the drivable area boundary that the URU can use to enter the drivable area by:

defining a plurality of hypotheses relating to the URU's trajectory for crossing the drivable area boundary; and

using the perception information and the plurality of hypotheses for identifying the plurality of entry points.

11. The computer program product of claim 10 , wherein the plurality of hypotheses comprise at least one of the following:

the URU will enter the drivable area from an entry point on the drivable area boundary that is closest to the URU's current position;

the URU will enter the drivable area from an entry point on the drivable area boundary where the drivable area boundary intersects with a heading vector of the URU; or

the URU will enter the drivable area from an entry point that is determined based on historical information about the drivable area.

12. The computer program product of claim 9 , further comprising programming instructions that cause the autonomous vehicle to predict, using the predicted trajectory, whether the URU will enter the drivable area.