Conflict resolver for a lidar data segmentation system of an autonomous vehicle

A method of controlling operation of an autonomous vehicle, the method executed by a computing device included on the autonomous vehicle, the method comprising:

• receiving light detection and ranging (lidar) data from a lidar sensor mounted on the autonomous vehicle, the lidar data comprising a plurality of points representative of positions of objects in a driving environment of the autonomous vehicle;
• responsive to receiving the lidar data, generating a plurality of segmentation scenes using a different respective segmentation algorithm to generate each of the segmentation scenes in the plurality of segmentation scenes, each of the segmentation algorithms in the plurality of segmentation algorithms utilizing the lidar data, wherein each segmentation scene in the plurality of segmentation scenes is representative of a same portion of the driving environment, and wherein each segmentation scene in the plurality of segmentation scenes comprises a plurality of regions, wherein each region in the plurality of regions for each segmentation scene in the plurality of segmentation scenes is indicative of an object in the objects, each region in a same segmentation scene being representative of a different object;
• generating a score for each segmentation scene in the plurality of segmentation scenes based upon a scoring function, wherein the scoring function outputs a first score for a first segmentation scene that comprises a first plurality of regions based upon assigning a sub-score to each region in the first plurality of regions based upon geometric features of the first plurality of regions;
• selecting a segmentation scene in the plurality of segmentation scenes based upon the scores;
• and
• controlling operation of the autonomous vehicle in the driving environment based upon the selected segmentation scene.

An autonomous vehicle comprising:

• a non-transitory computer-readable storage medium comprising instructions that, when executed by a processor of the autonomous vehicle, cause the processor to perform acts comprising:
• receiving a light detection and ranging (lidar) point cloud from a lidar sensor system mounted on the autonomous vehicle, the lidar point cloud comprising a plurality of points representative of positions of objects in a driving environment of the autonomous vehicle;
• responsive to receiving the lidar point cloud, generating a plurality of segmentation scenes using a different respective segmentation algorithm to generate each of the segmentation scenes in the plurality of segmentation scenes, each of the segmentation algorithms configured to generate a segmentation scene based upon the lidar point cloud, wherein each segmentation scene in the plurality of segmentation scenes comprises a plurality of regions, wherein each region in the plurality of regions for each segmentation scene in the plurality of segmentation scenes is indicative of an object in the objects, each region in a same segmentation scene being representative of a different object;
• generating a score for each segmentation scene in the plurality of segmentation scenes based upon a scoring function, wherein the scoring function outputs a first score for a first segmentation scene that comprises a first plurality of regions based upon assigning a sub-score to each region in the first plurality of regions based upon geometric characteristics of the first plurality of regions;
• selecting a segmentation scene in the plurality of segmentation scenes based upon the scores;
• and
• controlling at least one of an engine of the autonomous vehicle, a braking system of the autonomous vehicle, or a steering system of the autonomous vehicle during operation of the autonomous vehicle in the driving environment based upon the selected segmentation scene.