Point cloud segmentation using a coherent lidar for autonomous vehicle applications

1. A method comprising:

obtaining, by a sensing system of an autonomous vehicle (AV), a plurality of return points, each return point comprising one or more velocity values and one or more coordinates of a reflecting region that reflects a signal emitted by the sensing system, the one or more velocity values and the one or more coordinates obtained for a same first sensing frame;

identifying, in view of a distance in a coordinate-velocity space from each of a first set of the return points of the plurality of return points to at least one reference point in the coordinate-velocity space, that the first set of the return points is associated with a first object in an environment of the AV; and

causing a driving path of the AV to be determined in view of the first object.

2. The method of claim 1 , wherein the sensing system of the AV comprises a coherent light detection and ranging device (lidar) and wherein the signal emitted by the sensing system comprises a phase-modulated or a frequency-modulated electromagnetic wave.

3. The method of claim 2 , wherein the coherent lidar is a frequency-modulated continuous-wave lidar.

4. The method of claim 1 , wherein the one or more velocity values represent a component of a velocity of the reflecting region along a direction from a transmitter of the sensing system to the reflecting region.

5. The method of claim 1 , wherein the one or more velocity values comprise a first velocity value obtained using a first sensor of the sensing system and a second velocity value obtained using a second sensor of the sensing system, wherein the second sensor is located differently from the first sensor.

6. The method of claim 1 , wherein the plurality of return points belong to a sensing frame corresponding to one cycle of a transmitter of the sensing system.

7. The method of claim 1 , wherein identifying that the first set of the return points is associated with the first object comprises identifying the first set of the return points as a first cluster in the coordinate-velocity space, wherein identifying the first set of the return points as the cluster in the coordinate-velocity space comprises using a clustering criterion.

8. The method of claim 7 , wherein the clustering criterion is based, at least in part, on distances in the coordinate-velocity space from each of the first set of return points to the reference point in the coordinate-velocity space, the reference point being associated with a centroid of the first cluster.

9. The method of claim 1 , wherein identifying that the first set of the return points is associated with the first object comprises fitting the one or more velocity values and the one or more coordinates of each of the first set of the return points to a rigid-body equation characterizing a motion of a rigid body, wherein the rigid body is a body in which a distance between any points is fixed.

10. The method of claim 1 , further comprising:

identifying, in view of the one or more velocity values of a second set of the return points of the plurality of return points, that the second set of the return points is associated with a second object in the environment of the AV, wherein the second object is different from the first object.

11. The method of claim 10 , wherein identifying that the second set of the return points is associated with the second object comprises:

identifying the first set of the return points as a first cluster in the coordinate-velocity space; and

identifying the second set of the return points as a second cluster in the coordinate-velocity space, the second cluster centered differently from the first cluster.

12. The method of claim 10 , wherein identifying that the second set of the return points is associated with the second object comprises:

identifying a first set of fitting parameters for fitting the one or more velocity values and the one or more coordinates of each of the first set of the return points to a rigid-body equation;

identifying a second set of fitting parameters for fitting the one or more velocity values and the one or more coordinates of each of the first set of the return points to the rigid-body equation; and

determining that the second set of fitting parameters is different from the first set of fitting parameters.

13. The method of claim 10 , wherein identifying that the second set of the return points is associated with the second object comprises:

forming a first hypothesis that a combined set comprising the first set of the return points and the second set of the return points corresponds to a single object,

determining that a segmentation criterion for the combined set is not satisfied;

forming a second hypothesis that the second set of the return points corresponds to the second object; and

determining that the segmentation criterion for the second set of the return points is satisfied.

14. The method of claim 10 , wherein identifying that the first set of the return points is associated with the first object and that the second set of the return points is associated with the second object comprises:

identifying an initial set of the plurality of return points as a possible single object;

obtaining, by the sensing system, a plurality of additional return points, each additional return point comprising additional one or more velocity values and additional one or more coordinates of a respective reflecting region, the additional one or more velocity values and the additional one or more coordinates being obtained for a second sensing frame different from the first sensing frame; and

determining, based on the obtained plurality of additional return points, that the initial set comprises the first set of the return points corresponding to the first object and the second set of the return points corresponding to the second object.

15. A system comprising:

a sensing system of an autonomous vehicle (AV), the sensing system to:

obtain a plurality of return points, each return point comprising one or more velocity values and one or more coordinates of a reflecting region that reflects a signal emitted by the sensing system, the one or more velocity values and the one or more coordinates obtained for a same first sensing frame; and

a perception system of the AV, the perception system to:

identify, using a processing device of the perception system, in view of a distance in a coordinate-velocity space from each of a first set of the return points of the plurality of return points to at least one reference point in the coordinate-velocity space, that the first set of the return points is associated with a first object in an environment of the AV; and

cause a driving path of the AV to be determined in view of the first object.

16. The system of claim 15 , wherein the sensing system comprises a coherent light detection and ranging device (lidar) and wherein the signal emitted by the sensing system comprises a phase-modulated or a frequency-modulated electromagnetic wave.

17. The system of claim 15 , wherein the one or more velocity values comprise a first velocity value obtained using a first sensor of the sensing system and a second velocity value obtained using a second sensor of the sensing system, wherein the second sensor is located differently from the first sensor.

18. The system of claim 15 , wherein to identify that the first set of the return points is associated with the first object, the perception system is to:

identify the first set of the return points as a first cluster in the coordinate-velocity space, wherein identifying the first set of the return points as the cluster in the coordinate-velocity space comprises using a clustering criterion.

19. The system of claim 18 , wherein the clustering criterion is based, at least in part, on distances in the coordinate-velocity space from each of the first set of return points to the reference point in the coordinate-velocity space, the reference point being associated with a centroid of the first cluster.

20. A non-transitory computer-readable medium having instructions stored thereon that, when executed by a computing device, cause the computing device to:

obtain, from a sensing system of an autonomous vehicle (AV), a plurality of return points, each return point comprising one or more velocity values and one or more coordinates of a reflecting region that reflects a signal emitted by the sensing system, the one or more velocity values and the one or more coordinates obtained for a same first sensing frame;

identify, in view of a distance in a coordinate-velocity space from each of a first set of the return points of the plurality of return points to at least one reference point in the coordinate-velocity space, that the first set of the return points is associated with a first object in an environment of the AV; and

cause a driving path of the AV to be determined in view of the first object.