Systems and techniques for improved autonomous vehicle comfort and operation
Systems and techniques include functionality for improving a comfort, safety, and/or operation of an autonomous vehicle (AV) in an environment including a potential occlusion. An example method can include determining, by a computer of an AV and based on a first trajectory of an occluded hypothetical object, a dissipation boundary associated with the AV. The dissipation boundary includes a location where the AV is predicted to perceive the occluded hypothetical object given the first trajectory of the occluded hypothetical object; The method can include determining, based on the first trajectory of the occluded hypothetical object and a second trajectory of the AV, a hard brake boundary for the AV, and based on the dissipation boundary and the hard brake boundary, calibrating, by the computer, a speed of the AV.
1. A system comprising:
at least one memory; and
one or more processors coupled to the at least one memory, wherein the one or more processors are configured to:
determine, based on a first trajectory of an occluded hypothetical object, a dissipation boundary associated with an autonomous vehicle (AV), wherein the dissipation boundary comprises a location where the AV is predicted to perceive the occluded hypothetical object given the first trajectory of the occluded hypothetical object;
determine, based on the first trajectory of the occluded hypothetical object and a second trajectory of the AV, a hard brake boundary for the AV, the hard brake boundary indicating where the AV needs to have a decreased speed to avoid a collision with the occluded hypothetical object;
determine a time window having a duration based on a distance between the AV and the occluded hypothetical object, the duration based on a probability that the occluded hypothetical object exists; and
based on the dissipation boundary and the hard brake boundary, calibrate a speed of the AV, wherein the speed is calibrated so that the AV begins to decelerate at a selected time prior to reaching the hard brake boundary, the selected time based on the time window.
2. The system of claim 1 , wherein the occluded hypothetical object is occluded from a view of the AV by an object located within a proximity to the occluded hypothetical object, and wherein the AV is unable to detect the occluded hypothetical object prior to the dissipation boundary.
3. The system of claim 1 , wherein the hard brake boundary comprises a location where the AV needs to begin decreasing its speed to avoid a collision with the occluded hypothetical object at a location in space where the first trajectory of the occluded hypothetical object and the second trajectory of the AV are predicted to intersect.
4. The system of claim 1 , wherein the dissipation boundary is closer to the AV than the hard brake boundary, and wherein calibrating the speed of the AV comprises at least one of maintaining the speed of the AV and foregoing decreasing the speed of the AV before the hard brake boundary based on a determination that the dissipation boundary is closer to the AV than the hard brake boundary.
5. The system of claim 1 , wherein the hard brake boundary location is closer to the AV than the dissipation boundary location, and wherein calibrating the speed of the AV can include decreasing the speed of the AV before the hard brake boundary based on a determination that the hard brake boundary is closer to the AV than the dissipation boundary.
6. The system of claim 1 , wherein at least one of the first trajectory, the second trajectory, the dissipation boundary, and the hard brake boundary is estimated based on data from one or more sensors, the one or more sensors comprising at least one of a light detection and ranging (LIDAR) sensor, a radio detection and ranging (RADAR) sensor, a camera sensor, an inertial measurement unit (IMU), and an acoustic sensor.
7. The system of claim 6 , wherein the system comprises the AV, and wherein the AV is unable to detect the occluded hypothetical object prior to the dissipation boundary given at least one of a LIDAR coverage of the LIDAR sensor at one or more positions within the second trajectory of the AV, a RADAR coverage of the RADAR sensor at one or more positions within the second trajectory of the AV, and a camera sensor coverage of the camera sensor at one or more positions within the second trajectory of the AV.
8. A method comprising:
determining, by a computer of an autonomous vehicle (AV) and based on a first trajectory of an occluded hypothetical object, a dissipation boundary associated with the AV, wherein the dissipation boundary comprises a location where the AV is predicted to perceive the occluded hypothetical object given the first trajectory of the occluded hypothetical object;
determining, based on the first trajectory of the occluded hypothetical object and a second trajectory of the AV, a hard brake boundary for the AV, the hard brake boundary indicating where the AV needs to have a decreased speed to avoid a collision with the occluded hypothetical object;
determining a time window having a duration based on a distance between the AV and the occluded hypothetical object, the duration based on a probability that the occluded hypothetical object exists; and
based on the dissipation boundary and the hard brake boundary, calibrating, by the computer, a speed of the AV, wherein the speed is calibrated so that the AV begins to decelerate at a selected time prior to reaching the hard brake boundary, the selected time based on the time window.
9. The method of claim 8 , wherein the occluded hypothetical object is occluded from a view of the AV by an object located within a proximity to the occluded hypothetical object, and wherein the AV is unable to detect the occluded hypothetical object prior to the dissipation boundary.
10. The method of claim 8 , wherein the hard brake boundary comprises a location where the AV needs to begin decreasing its speed to avoid a collision with the occluded hypothetical object at a location in space where the first trajectory of the occluded hypothetical object and the second trajectory of the AV are predicted to intersect.
11. The method of claim 8 , wherein the dissipation boundary is closer to the AV than the hard brake boundary, and wherein calibrating the speed of the AV comprises at least one of maintaining the speed of the AV and foregoing decreasing the speed of the AV before the hard brake boundary based on a determination that the dissipation boundary is closer to the AV than the hard brake boundary.
12. The method of claim 8 , wherein the hard brake boundary location is closer to the AV than the dissipation boundary location, and wherein calibrating the speed of the AV comprises decreasing the speed of the AV before the hard brake boundary based on a determination that the hard brake boundary is closer to the AV than the dissipation boundary.
13. The method of claim 8 , wherein at least one of the first trajectory, the second trajectory, the dissipation boundary, and the hard brake boundary is estimated based on data from one or more sensors, the one or more sensors comprising at least one of a light detection and ranging (LIDAR) sensor, a radio detection and ranging (RADAR) sensor, a camera sensor, an inertial measurement unit (IMU), and an acoustic sensor.
14. The method of claim 13 , wherein the AV is unable to detect the occluded hypothetical object prior to the dissipation boundary given at least one of a LIDAR coverage of the LIDAR sensor at one or more positions within the second trajectory of the AV, a RADAR coverage of the RADAR sensor at one or more positions within the second trajectory of the AV, and a camera sensor coverage of the camera sensor at one or more positions within the second trajectory of the AV.
15. A non-transitory computer-readable storage medium comprising instructions which, when executed by one or more processors, cause the one or more processors to:
determine, based on a first trajectory of an occluded hypothetical object, a dissipation boundary associated with an autonomous vehicle (AV), wherein the dissipation boundary comprises a location where the AV is predicted to perceive the occluded hypothetical object given the first trajectory of the occluded hypothetical object;
determine, based on the first trajectory of the occluded hypothetical object and a second trajectory of the AV, a hard brake boundary for the AV, the hard brake boundary indicating where the AV needs to have a decreased speed to avoid a collision with the occluded hypothetical object;
determine a time window having a duration based on a distance between the AV and the occluded hypothetical object, the duration based on a probability that the occluded hypothetical object exists; and
based on the dissipation boundary and the hard brake boundary, calibrate a speed of the AV, wherein the speed is calibrated so that the AV begins to decelerate at a selected time prior to reaching the hard brake boundary, the selected time based on the time window.
16. The non-transitory computer-readable storage medium of claim 15 , wherein the occluded hypothetical object is occluded from a view of the AV by an object located within a proximity to the occluded hypothetical object, and wherein the AV is unable to detect the occluded hypothetical object prior to the dissipation boundary.
17. The non-transitory computer-readable storage medium of claim 15 , wherein the hard brake boundary comprises a location where the AV needs to begin decreasing its speed to avoid a collision with the occluded hypothetical object at a location in space where the first trajectory of the occluded hypothetical object and the second trajectory of the AV are predicted to intersect.
18. The non-transitory computer-readable storage medium of claim 15 , wherein the dissipation boundary is closer to the AV than the hard brake boundary, and wherein calibrating the speed of the AV comprises at least one of maintaining the speed of the AV and foregoing decreasing the speed of the AV before the hard brake boundary based on a determination that the dissipation boundary is closer to the AV than the hard brake boundary.
19. The non-transitory computer-readable storage medium of claim 15 , wherein the hard brake boundary location is closer to the AV than the dissipation boundary location, and wherein calibrating the speed of the AV comprises decreasing the speed of the AV before the hard brake boundary based on a determination that the hard brake boundary is closer to the AV than the dissipation boundary.
20. The non-transitory computer-readable storage medium of claim 15 , wherein at least one of the first trajectory, the second trajectory, the dissipation boundary, and the hard brake boundary is estimated based on data from one or more sensors, the one or more sensors comprising at least one of a light detection and ranging (LIDAR) sensor, a radio detection and ranging (RADAR) sensor, a camera sensor, an inertial measurement unit (IMU), and an acoustic sensor.