Eliminating sensor self-hit data

Abstract

The subject disclosure relates to ways to identify self-hit data collected by autonomous vehicle (AV) sensors. In some aspects, a method of the disclosed technology includes steps for generating a geometric model of an autonomous vehicle (AV), wherein the geometric model specifies physical boundaries of the AV in three-dimensional (3D) space, collecting sensor data for an environment around the AV, and identifying one or more data points, from among the collected sensor data, that correspond with a surface of the AV. Systems and machine-readable media are also provided.

What is claimed is: 
     
         1 . A computer-implemented method comprising:
 accessing a geometric model that approximates shapes and locations of surfaces on an autonomous vehicle (AV), wherein the geometric model is configured to be applied to a varying geometry of the AV;   modifying the geometric model to generate a modified geometric model that is specific to a first geometry of the AV;   generating a filter based on the modified geometric model;   accessing sensor data for an environment around the AV; and   applying the filter to the sensor data to identify a surface of the AV in the first geometry of the AV from the sensor data for the environment around the AV.   
     
     
         2 . The computer-implemented method of  claim 1 , further comprising:
 accessing the geometric model;   modifying the geometric model to generate another modified geometric model that is specific to a second geometry of the AV;   generating another filter based on the another modified geometric model; and   applying the another filter to the sensor data to identify a surface of the AV in the second geometry of the AV from the sensor data for the environment around the AV.   
     
     
         3 . The computer-implemented method of  claim 1 , wherein the geometric model comprises one or more error buffers that are tunable to generate the modified geometric model. 
     
     
         4 . The computer-implemented method of  claim 3 , wherein the one or more error buffers are tunable to account for the varying geometry of the AV. 
     
     
         5 . The computer-implemented method of  claim 1 , wherein the filter is a matrix configured to perform transformations on arrays or vectors of the sensor data. 
     
     
         6 . The computer-implemented method of  claim 1 , wherein the filter is generated based on a geometric coordinate system of the modified geometric model. 
     
     
         7 . The computer-implemented method of  claim 1 , wherein applying the filter to the sensor data to identify the surface of the AV in the first geometry of the AV further comprises:
 determining that a subset of the sensor data that corresponds to the surface of the AV in the first geometry of the AV resides within the modified geometric model; and   identifying that the surface of the AV in the first geometry of the AV is in the sensor data based on the subset of the sensor data residing within the modified geometric model.   
     
     
         8 . The computer-implemented method of  claim 1 , wherein the modified geometric model is adaptable across a plurality of coordinate systems. 
     
     
         9 . The computer-implemented method of  claim 1 , further comprising facilitating navigation of the AV based on the surface of the AV in the first geometry of the AV being identified from the sensor data for the environment around the AV. 
     
     
         10 . A system comprising:
 one or more processors; and   at least one computer-readable storage medium having stored therein instructions which, when executed by the one or more processors, cause the one or more processors to:
 access a geometric model that approximates shapes and locations of surfaces on an autonomous vehicle (AV), wherein the geometric model is configured to be applied to a varying geometry of the AV; 
 modify the geometric model to generate a modified geometric model that is specific to a first geometry of the AV; 
 generate a filter based on the modified geometric model; 
 access sensor data for an environment around the AV; and 
 apply the filter to the sensor data to identify a surface of the AV in the first geometry of the AV from the sensor data for the environment around the AV. 
   
     
     
         11 . The system of  claim 10 , wherein the instructions further cause the one or more processors to:
 access the geometric model;   modify the geometric model to generate another modified geometric model that is specific to a second geometry of the AV;   generate another filter based on the another modified geometric model; and   apply the another filter to the sensor data to identify a surface of the AV in the second geometry of the AV from the sensor data for the environment around the AV.   
     
     
         12 . The system of  claim 10 , wherein the geometric model comprises one or more error buffers that are tunable to generate the modified geometric model. 
     
     
         13 . The system of  claim 12 , wherein the one or more error buffers are tunable to account for the varying geometry of the AV. 
     
     
         14 . The system of  claim 10 , wherein the filter is a matrix configured to perform transformations on arrays or vectors of the sensor data. 
     
     
         15 . The system of  claim 10 , wherein the filter is generated based on a geometric coordinate system of the modified geometric model. 
     
     
         16 . The system of  claim 10 , wherein the instructions further cause the one or more processors to:
 determine that a subset of the sensor data that corresponds to the surface of the AV in the first geometry of the AV resides within the modified geometric model; and   identify that the surface of the AV in the first geometry of the AV is in the sensor data based on the subset of the sensor data residing within the modified geometric model.   
     
     
         17 . The system of  claim 10 , wherein the modified geometric model is adaptable across a plurality of coordinate systems. 
     
     
         18 . The system of  claim 10 , the instructions further cause the one or more processors to facilitate navigation of the AV based on the surface of the AV in the first geometry of the AV being identified from the sensor data for the environment around the AV. 
     
     
         19 . A non-transitory computer-readable storage medium storing instructions for causing one or more processors to:
 access a geometric model that approximates shapes and locations of surfaces on an autonomous vehicle (AV), wherein the geometric model is configured to be applied to a varying geometry of the AV;   modify the geometric model to generate a modified geometric model that is specific to a first geometry of the AV;   generate a filter based on the modified geometric model;   access sensor data for an environment around the AV; and   apply the filter to the sensor data to identify a surface of the AV in the first geometry of the AV from the sensor data for the environment around the AV.   
     
     
         20 . The non-transitory computer-readable storage medium of  claim 19 , wherein the instructions further cause the one or more processors to:
 access the geometric model;   modify the geometric model to generate another modified geometric model that is specific to a second geometry of the AV;   generate another filter based on the another modified geometric model; and   apply the another filter to the sensor data to identify a surface of the AV in the second geometry of the AV from the sensor data for the environment around the AV.