Driving difficulty heat maps for autonomous vehicles

1. A method of generating a driving difficulty heat map for autonomous vehicles, the method comprising:

inputting, by one or more processors, log data generated by a vehicle being driven in a manual driving mode for a segment of a route into a disengage model in order to generate an output identifying a likelihood of a vehicle driving in an autonomous driving mode requiring a disengage from the autonomous driving mode along the segment of the route, the log data being collected within a geographic area;

generating, by the one or more processors, a grid for the geographic area, the grid including a plurality of cells;

assigning, by the one or more processors, the output to one of the plurality of cells;

using, by the one or more processors, the plurality of cells and assigned output to generate a driving difficulty heat map for the geographic area; and

identifying, by the one or more processors, a geographic area to be excluded from a service area for the autonomous vehicles in real time based on the driving difficulty heat map.

2. The method of claim 1 , wherein dimensions of the cells are selected based on a length of the segment of the route.

3. The method of claim 1 , wherein the disengage model is a machine learned model trained using log data collected by a vehicle driving in an autonomous driving mode in a second geographic area different from the geographic area.

4. The method of claim 3 , further comprising training the disengage model using a number of different types of maneuvers from the log data.

5. The method of claim 3 , further comprising training the disengage model using a number of different types of maneuvers per a predetermined driving distance from the log data.

6. The method of claim 1 , wherein the output further includes a number of likely disengages along the segment of the route.

7. The method of claim 1 , wherein the output further includes a lane-level location of any potential disengages along that segment.

8. A system for generating a driving difficulty heat map for autonomous vehicles, the system comprising one or more processors configured to:

input log data generated by a vehicle being driven in a manual driving mode for a segment of a route into a disengage model in order to generate an output identifying a likelihood of a vehicle driving in an autonomous driving mode requiring a disengage from the autonomous driving mode along the segment of the route, the log data being collected within a geographic area;

generate a grid for the geographic area, the grid including a plurality of cells;

assign the output to one of the plurality of cells; and

use the plurality of cells and assigned output to generate the driving difficulty heat map for the geographic area and

identifying, by the one or more processors, a geographic area to be excluded from a service area for the autonomous vehicles in real time based on the driving difficulty heatmap.

9. The system of claim 8 , wherein dimensions of the cells are selected based on a length of the segment of the route.

10. The system of claim 8 , wherein the disengage model is a machine learned model trained using log data collected by a vehicle driving in an autonomous driving mode in a second geographic area different from the geographic area.

11. The system of claim 10 , wherein the one or more processors are further configured to train the disengage model using a number of different types of maneuvers from the log data.

12. The system of claim 10 , wherein the one or more processors are further configured to train the disengage model using a number of different types of maneuvers per a predetermined driving distance from the log data.

13. The system of claim 8 , wherein the output further includes a number of likely disengages along the segment of the route.

14. The system of claim 8 , wherein the output further includes a lane-level location of any potential disengages along that segment.

15. A non-transitory computing device-readable medium on which instructions are stored, the instructions, when executed by one or more processors, cause the one or more processors to perform a method of generating a driving difficulty heat map for autonomous vehicles, the method comprising:

inputting log data generated by a vehicle being driven in a manual driving mode for a segment of a route into a disengage model in order to generate an output identifying a likelihood of a vehicle driving in an autonomous driving mode requiring a disengage from the autonomous driving mode along the segment of the route, the log data being collected within a geographic area;

generating a grid for the geographic area, the grid including a plurality of cells;

assigning the output to one of the plurality of cells; and

using the plurality of cells and assigned output to generate the driving difficulty heat map for the geographic area; and

identifying a geographic area to be excluded from a service area for the autonomous vehicles in real time based on the driving difficulty heat map.

16. The medium of claim 15 , wherein dimensions of the cells are selected based on a length of the segment of the route.

17. The medium of claim 15 , wherein the disengage model is a machine learned model trained using log data collected by a vehicle driving in an autonomous driving mode in a second geographic area different from the geographic area.

18. The medium of claim 17 , wherein the method further comprises training the disengage model using a number of different types of maneuvers from the log data.

19. The medium of claim 17 , wherein the method further comprises training the disengage model using a number of different types of maneuvers per a predetermined driving distance from the log data.

20. The medium of claim 15 , wherein the output further includes a number of likely disengages along the segment of the route.