Using high definition maps for generating synthetic sensor data for autonomous vehicles

Abstract

According to an aspect of an embodiment, operations may comprise accessing high definition (HD) map data of a region, presenting, via a user interface, information describing the HD map data, receiving instructions, via the user interface, for modifying the HD map data by adding one or more synthetic objects to locations in the HD map data, modifying the HD map data based on the received instructions, and generating a synthetic track in the modified HD map data comprising, for each of one or more vehicle poses, generated synthetic sensor data based on the one or more synthetic objects in the modified HD map data.

What is claimed is: 
     
         1 . A method, comprising:
 receiving instructions for adding one or more synthetic objects to one or more locations in a map;   based at least on the instructions, generating a modified map based at least on including the one or more synthetic objects at the one or more locations in the map;   based at least on the instructions, generating synthetic track data using the modified map, the synthetic track data corresponding to a simulated trajectory of a simulated machine and including synthetic sensor data corresponding to the one or more synthetic objects; and   performing one or more operations based at least on the synthetic track data.   
     
     
         2 . The method of  claim 1 , wherein the one or more operations include evaluating performance of the simulated machine based at least on the synthetic track data. 
     
     
         3 . The method of  claim 1 , wherein the generating the synthetic track data is performed within a simulation environment generated based at least on the modified map. 
     
     
         4 . The method of  claim 1 , wherein the one or more operations include using the synthetic track data as training data for updating one or more parameters of one or more neural networks. 
     
     
         5 . The method of  claim 4 , wherein the one or more neural networks, after updating, are deployed in one or more real-world machines for use in performing one or more navigation operations. 
     
     
         6 . The method of  claim 1 , wherein the modifying of the map includes one or more of:
 adding respective point cloud representations corresponding to the one or more synthetic objects to point cloud data corresponding to the map; or   updating landmark map data corresponding to the map to include respective location information of the one or more synthetic objects.   
     
     
         7 . A processor comprising:
 processing circuitry to cause performance of operations comprising:
 receiving instructions for adding one or more synthetic objects to one or more locations in a map; 
 based at least on the instructions, generating a modified map based at least on including the one or more synthetic objects at the one or more locations in the map; 
 generating a simulated environment using the modified map; 
 generating, based at least on a simulated trajectory of a simulated machine within the simulation environment, synthetic track data corresponding to the simulated trajectory and including synthetic sensor data corresponding to the one or more synthetic objects; and 
 performing one or more operations based at least on the synthetic track data. 
   
     
     
         8 . The processor of  claim 7 , wherein the instructions further define one or more of:
 the simulated trajectory;   one or more lane elements corresponding to the simulated trajectory; or   one or more waypoints constraining the simulated trajectory.   
     
     
         9 . The processor of  claim 7 , wherein the one or more operations include evaluating performance of the simulated machine based at least on the synthetic track data. 
     
     
         10 . The processor of  claim 7 , wherein the synthetic track data includes simulated sensor data generated using the simulated machine. 
     
     
         11 . The processor of  claim 10 , wherein the one or more operations include using the simulated sensor data as training data to update one or more parameters of one or more neural networks. 
     
     
         12 . The processor of  claim 11 , wherein the one or more neural networks, after updating, are deployed in one or more real-world machines for use in performing one or more navigation or control operations. 
     
     
         13 . The processor of  claim 7 , wherein the modifying of the map includes one or more of:
 adding respective point cloud representations corresponding to the one or more synthetic objects to point cloud data corresponding to the map; or   updating landmark map data corresponding to the map to include respective location information of the one or more synthetic objects.   
     
     
         14 . A system comprising:
 one or more processing units to perform operations comprising:
 receiving instructions for adding one or more synthetic objects to one or more locations in a map; 
 based at least on the instructions, generating a modified map based at least on including the one or more synthetic objects at the one or more locations; 
 generating synthetic track data using the modified map, the synthetic track data corresponding to a simulated trajectory for a simulated machine and including synthetic sensor data corresponding to the one or more synthetic objects in the modified map; and 
 performing one or more operations based at least on the synthetic track data. 
   
     
     
         15 . The system of  claim 14 , wherein the instructions further define one or more of:
 the simulated trajectory;   one or more lane elements corresponding to the simulated trajectory; or   one or more waypoints constraining the simulated trajectory.   
     
     
         16 . The system of  claim 14 , wherein the one or more operations include evaluating performance of the simulated machine based at least on the synthetic track data. 
     
     
         17 . The system of  claim 14 , wherein the generating the synthetic track data is performed using a simulation environment generated based at least on the modified map. 
     
     
         18 . The system of  claim 14 , wherein the one or more operations include using the synthetic track data as training data to update one or more parameters of one or more neural networks. 
     
     
         19 . The system of  claim 18 , wherein the one or more neural networks, after updating, are deployed in one or more real-world machine for use in performing one or more navigation operations. 
     
     
         20 . The system of  claim 14 , wherein the modifying of the map includes one or more of:
 adding respective point cloud representations corresponding to the one or more synthetic objects to point cloud data corresponding to the map; or   updating landmark map data corresponding to the map to include respective location information of the one or more synthetic objects.