Lidar based detection of road surface features

Abstract

An AV is described herein. The AV includes a lidar sensor system. The AV additionally includes a computing system that executes a road surface analysis component to determine, based upon lidar sensor data, whether a road surface feature is present on or in a roadway in a travel path of the AV. The AV can be configured to initiate a mitigation maneuver responsive to determining that the road surface feature is present. Performing the mitigation maneuver causes the AV to avoid the road surface feature or decelerate prior to reaching the road surface feature, thereby improving the apparent quality or comfort of the ride to a passenger of the AV.

What is claimed is: 
     
         1 . An autonomous vehicle (AV) comprising:
 a lidar sensor system; and   a computing system that is in communication with the lidar sensor system, wherein the computing system comprises:
 a processor; and 
 memory that stores instructions that, when executed by the processor, cause the processor to perform acts comprising:
 receiving lidar data output by the lidar sensor system, the lidar data comprising a plurality of points, the points indicative of positions of objects in a driving environment of the AV; 
 identifying a subset of the points that are representative of a road surface in a travel path of the AV; and 
 outputting, based upon the identified subset of the points, an indication that a road surface feature is likely present on or in the road surface that is in the travel path of the AV. 
 
   
     
     
         2 . The AV of  claim 1 , wherein the acts further comprise:
 controlling the AV to avoid the road surface feature or to change a speed of the AV prior to traveling over the road surface feature.   
     
     
         3 . The AV of  claim 1 , wherein determining that the road surface feature is likely present on or in the road surface comprises:
 computing, based upon the identified subset of the points, a probability that the road surface feature is present on or in the road surface; and   determining that the road surface feature is likely present based upon the computed probability exceeding a threshold.   
     
     
         4 . The AV of  claim 1 , wherein determining that the road surface feature is likely present comprises:
 comparing the identified subset of the points to a lidar signature of a known road surface feature; and   based upon the comparing, determining that the identified subset of the points matches the lidar signature.   
     
     
         5 . The AV of  claim 4 , wherein the lidar signature of the known road surface feature comprises a plurality of lidar points that are representative of the known road surface feature. 
     
     
         6 . The AV of  claim 4 , wherein comparing the identified subset of the points to the lidar signature of the known road surface feature comprises computing a similarity score between the lidar signature and the identified subset of the points, and wherein determining that the identified subset of the points matches the lidar signature comprises determining that the similarity score exceeds a threshold value. 
     
     
         7 . The AV of  claim 6 , wherein computing the similarity score comprises:
 selecting a pair of points, wherein one of the pair of points is drawn from the identified subset of the points and the other of the pair of points is drawn from the lidar signature; and   computing a distance between the pair of points, wherein the similarity score is based upon the computed distance.   
     
     
         8 . The AV of  claim 1 , wherein determining that the road surface feature is likely present comprises:
 comparing the identified subset of the points to a plurality of lidar signatures that are representative of known road surface features, wherein the lidar signatures comprise:
 a lidar signature of a manhole cover; 
 a lidar signature of a pothole; 
 a lidar signature of a speedbump; or 
 a lidar signature of motor vehicle accident debris. 
   
     
     
         9 . The AV of  claim 1 , wherein determining that the road surface feature is likely present comprises:
 identifying a conflict between the identified subset of the points and a height map that is representative of the road surface; and   determining that the road surface feature is likely present in or on the road surface based upon the conflict being identified between the identified subset of the points and the height map.   
     
     
         10 . The AV of  claim 1 , wherein determining that the road surface feature is likely present comprises:
 generating a first height map of the road surface based upon the identified subset of the points;   determining that a conflict exists between the first height map of the road surface and a predefined height map of the road surface; and   determining that the road surface feature is likely present based upon determining that the conflict exists.   
     
     
         11 . The AV of  claim 10 , wherein determining that the road surface feature is likely present comprises determining that a number of conflicts between the first height map and the predefined height map exceeds a threshold. 
     
     
         12 . A method, comprising:
 generating a point cloud by way of a lidar sensor system, the point cloud comprising points that are indicative of positions of objects in an environment;   filtering the point cloud to identify a group of the points that are representative of a road surface in a path of a vehicle;   determining, based upon the identified group of the points, that a road surface feature is likely present on or in the road surface; and   outputting an indication of the presence of the road surface feature on or in the road surface.   
     
     
         13 . The method of  claim 12 , wherein the group of the points that is representative of the road surface is a first group of the points, wherein filtering the point cloud to identify the first group of the points comprises:
 identifying a second group of the points that is representative of an object in the environment other than the road surface; and   excluding the second group of the points from the first group of the points.   
     
     
         14 . The method of  claim 12 , wherein filtering the point cloud to identify the group of the points that are representative of the road surface is based upon map data that indicates positions of known roadways in the environment. 
     
     
         15 . The method of  claim 12 , further comprising illuminating a portion of the road surface in the path of the vehicle by way of the lidar sensor system such that a portion of the point cloud that represents the portion of the road surface has a density of at least 1000 points per square meter. 
     
     
         16 . The method of  claim 12 , wherein determining that the road surface feature is likely present on or in the road surface comprises computing a probability that the road surface feature is present on or in the road surface. 
     
     
         17 . The method of  claim 16 , wherein outputting the indication of the presence of the road surface feature comprises outputting an indication of the computed probability, the method further comprising:
 responsive to the computed probability exceeding a first threshold, causing the vehicle to avoid the road surface feature or to decelerate prior to traveling over the road surface feature.   
     
     
         18 . The method of  claim 17 , the method further comprising:
 responsive to the computed probability exceeding the first threshold and a second threshold, causing the vehicle to avoid the road surface feature; and   responsive to the computed probability exceeding the first threshold and failing to exceed the second threshold, causing the vehicle to decelerate prior to traveling over the road surface feature.   
     
     
         19 . The method of  claim 12 , further comprising:
 determining a type of the road surface feature that is likely present on or in the road surface;   responsive to determining that the type of the road surface feature is a first type, causing the vehicle to avoid the road surface feature; and   responsive to determining that the type of the road surface feature is a second type, causing the vehicle to decelerate prior to traveling over the road surface feature.   
     
     
         20 . A computer-readable storage medium that stores instructions that, when executed by a processor, cause the processor to perform acts comprising:
 receiving a point cloud generated by way of a lidar sensor system, the point cloud comprising points that are indicative of positions of objects in a driving environment of a vehicle;   excluding a first group of the points that are representative of a first object from the point cloud to generate a second group of the points that are representative of a road surface in a path of a vehicle;   determining, based upon the second group of the points, that a road surface feature is likely present on or in the road surface; and   responsive to determining that the road surface feature is likely present on or in the road surface, controlling the vehicle to avoid the road surface feature or to change a speed of the vehicle prior to the vehicle traveling over the road surface feature.