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

1. A method, comprising:

accessing high definition (HD) map data of a region, the HD map data including landmark map data and occupancy map data, the landmark map data including location information related to one or more objects included in the HD map data, the occupancy map data including point cloud data representing the region and obtained from one or more light detection and ranging (LIDAR) sensors;

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 and for defining one or more of: a route of a vehicle, one or more lane elements corresponding to the route, or one or more waypoints constraining the route;

modifying the HD map data based at least on the instructions, wherein the modifying of the HD map data includes:

adding respective point cloud representations of the one or more synthetic objects to the point cloud data of the occupancy map data; and

updating the landmark map data to include respective location information of the one or more synthetic objects; and

generating synthetic track data in the modified HD map data, the synthetic track data comprising, for one or more vehicle poses, synthetic sensor data based at least on the one or more synthetic objects in the modified HD map data such that the synthetic track data corresponds to a simulated vehicle trajectory within the region, the simulated vehicle trajectory being determined based at least on the instructions.

2. The method of claim 1 , wherein the synthetic track data simulates a lane closure without an actual lane closure in the real world to allow for testing of navigation of an autonomous vehicle along the simulated vehicle trajectory that corresponds to the synthetic track data.

3. The method of claim 1 , wherein the instructions for modifying the HD map data further comprise one or more of:

removing a synthetic object from a location in the HD map data; or

moving a synthetic object from a first location in the HD map data to a second location in the HD map data.

4. The method of claim 1 , further comprising training a deep learning model based at least on the synthetic track data, at least one instance of the deep learning model used by an autonomous vehicle for navigation along a route.

5. The method of claim 1 , wherein the one or more synthetic objects comprise one or more of: a synthetic traffic sign, a synthetic traffic cone, a synthetic traffic light, a synthetic lane line, a synthetic curb, a synthetic barrier, or a synthetic dynamic object.

6. The method of claim 1 , wherein the synthetic sensor data comprises one or more of: generated synthetic LIDAR data, or generated synthetic camera data.

7. A processor comprising:

processing circuitry to cause performance of operations comprising:

receiving, via a user interface and based at least on a presentation of information corresponding to HD map data of a region, one or more first inputs corresponding to a modification of the HD map data by adding one or more synthetic objects to locations in the HD map data and one or more second inputs corresponding to one or more of: a route of a vehicle, one or more lane elements corresponding to the route, or one or more waypoints constraining the route;

modifying the HD map data based at least on the one or more first inputs, wherein the modifying of the HD map data includes:

adding respective point cloud representations of the one or more synthetic objects to occupancy map data of the HD map data, and

updating landmark map data of the HD map data to include location information related to the one or more synthetic objects in response to adding the respective point cloud representations to the occupancy map data; and

generating, based at least on the one or more first inputs and the one or more second inputs, synthetic track data corresponding to the modified HD map data and corresponding to the route, the synthetic track data comprising, for one or more vehicle poses, synthetic sensor data generated based at least on the one or more synthetic objects in the modified HD map data, the synthetic track data corresponding to a simulated vehicle trajectory within the region.

8. The processor of claim 7 , wherein the synthetic track data simulates a lane closure without an actual lane closure in the real world to allow for testing of one or more navigation operations associated with an autonomous vehicle along the simulated vehicle trajectory that corresponds to the synthetic track data.

9. The processor of claim 7 , wherein the one or more first inputs further correspond to one or more of:

removing a synthetic object from a location in the HD map data; or

moving a synthetic object from a first location in the HD map data to a second location in the HD map data.

10. The processor of claim 7 , further comprising training a deep learning model based at least on the synthetic track data, at least one instance of the deep learning model used in an autonomous vehicle for navigation.

11. The processor of claim 7 , wherein the one or more synthetic objects comprise one or more of: a synthetic traffic sign, a synthetic traffic cone, a synthetic traffic light, a synthetic lane line, a synthetic curb, a synthetic barrier, or a synthetic dynamic object.

12. The processor of claim 7 , wherein the synthetic sensor data comprises one or more of: generated synthetic LIDAR data, or generated synthetic camera data.

13. A system comprising:

one or more processing units to perform operations comprising:

receiving, via a user interface and based at least on a presentation of information corresponding to HD map data of a region, one or more first inputs corresponding to a modification of the HD map data by adding one or more synthetic objects to locations in the HD map data and one or more second inputs corresponding to a route of a vehicle;

modifying the HD map data based at least on the one or more first inputs, wherein the modifying of the HD map data includes:

updating, with respective representations of the one or more synthetic objects, occupancy map data of the HD map data that includes point cloud data representing the region, and

updating landmark map data of the HD map data to include location information related to the one or more synthetic objects; and

generating, based at least on the one or more first inputs and the one or more second inputs, synthetic track data corresponding to the modified HD map data and corresponding to the route, the synthetic track data comprising, for one or more vehicle poses, synthetic sensor data generated based at least on the one or more synthetic objects in the modified HD map data, the synthetic track data corresponding to a simulated vehicle trajectory within the region.

14. The system of claim 13 , wherein the synthetic track data simulates a lane closure without an actual lane closure in the real world to allow for testing of one or more navigation operations associated with an autonomous vehicle along the simulated vehicle trajectory that corresponds to the synthetic track data.

15. The system of claim 13 , wherein the one or more first inputs further correspond to removing a synthetic object from a location in the HD map data.

16. The system of claim 13 , wherein the one or more first inputs further correspond to moving a synthetic object from a first location in the HD map data to a second location in the HD map data.

17. The system of claim 13 , further comprising training a deep learning model based at least on the synthetic track data, at least one instance of the deep learning model used in an autonomous vehicle for navigation.

18. The system of claim 13 , wherein the one or more synthetic objects comprise one or more of: a synthetic traffic sign, a synthetic traffic cone, a synthetic traffic light, a synthetic lane line, a synthetic curb, a synthetic barrier, or a synthetic dynamic object.

19. The system of claim 13 , wherein the synthetic sensor data comprises one or more of: generated synthetic LIDAR data, or generated synthetic camera data.

20. The system of claim 13 , wherein the one or more second inputs corresponding to the route include one or more of: one or more lane elements corresponding to the route or one or more waypoints constraining the route.