Perception system for autonomous vehicle

1. A computer-implemented method for operating a vehicle, the method comprising:

obtaining stored data identifying a set of static objects which are present in an area of a road segment in association with map data indicative of the road segment;

autonomously operating the vehicle to travel on the road segment by:

determining a route for the vehicle to use for travelling for a transport service, wherein the route comprises the road segment;

collecting current sensor data of a scene that includes the area of the road segment, using a sensor set of the vehicle;

determining a plurality of perceived objects of the scene based at least in part on the current sensor data;

determining at least one perceived object that is present in the area of the road segment is a non-static object by subtracting the set of static objects which are present in the area of the road segment from the plurality of perceived objects of the scene;

tracking the at least one perceived object as the vehicle progresses on the road segment, using the sensor set, without tracking any object in the set of static objects by excluding current sensor data associated with any object in the set of static objects for the road segment; and

determining a trajectory for the vehicle to follow based, at least in part, on the route and the at least one perceived object.

2. The computer-implemented method of claim 1 , wherein collecting the current sensor data includes collecting current image data from a camera set of the vehicle.

3. The computer-implemented method of claim 2 , wherein the current image data is recorded from one or more sets of stereoscopic cameras.

4. The computer-implemented method of claim 2 , wherein the stored data comprises stored image data, and wherein determining that the at least one perceived object that is present in the area of the road segment is a non-static object comprises:

comparing the current image data collected from the camera set with the stored image data that identifies the set of static objects in order to determine portions of the current image data which depict one or more of the static objects.

5. The computer-implemented method of claim 4 , wherein determining that the at least one perceived object that is present in the area of the road segment is a non-static object comprises includes ignoring or eliminating the determined portions of the current image data which correspond to the set of static objects.

6. The computer-implemented method of claim 4 , wherein comparing the current image data with the stored image data comprises:

performing a transformation on one of the current image data or the stored image data to account for a variation in lighting as between the current image data and the stored image data.

7. The computer-implemented method of claim 6 , wherein performing the transformation includes altering one or more pixel values of at least one of the current image data or the stored image data for hue, contrast, or brightness.

8. The computer-implemented method of claim 6 , further comprising:

determining a probability that the at least one perceived object will intersect a path of the vehicle or collide with the vehicle.

9. The computer-implemented method of claim 8 , wherein determining the probability is based on a worst-case scenario for the at least one perceived object.

10. The computer-implemented method of claim 6 , wherein performing the transformation includes selecting a transformation based on contextual input that includes at least one of a weather input, day of year input, or time of day input.

11. The computer-implemented method of claim 1 , wherein tracking the at least one perceived object includes determining a trajectory of the at least one perceived object.

12. The computer-implemented method of claim 1 , wherein the stored data includes a set of submaps, each submap of the set of submaps correlating to a respective area of a respective road segment on the route for the vehicle to use on the trip, each submap of the set of submaps including data that identifies a corresponding set of static objects present in the respective area of a the respective road segment for that submap.

13. The computer-implemented method of claim 12 , further comprising:

retrieving the set of submaps based on a determined location of the autonomous vehicle.

14. A computer system comprising:

a memory to store a set of instructions; and

one or more processors to use the instructions to:

obtain stored data that identifies a set of static objects which are present in an area of a road segment in association with map data indicative of the road segment; and

autonomously operate a vehicle to travel on the road segment by:

determining a route for the vehicle to use for travelling for a transport service, wherein the route comprises the road segment

collecting current sensor data of a scene that includes the area of the road segment, using a sensor set of the vehicle;

determining a plurality of perceived objects of the scene based at least in part on the current sensor data;

determining that at least one perceived object that is present in the area of the road segment is a non-static object by subtracting the set of static objects which are present in the area of the road segment from the plurality of perceived objects of the scene;

tracking the at least one perceived object as the vehicle progresses on the road segment, using the sensor set, without tracking any object in the set of static objects by excluding sensor data associated with any object in the set of static objects for the road segment; and

determining a trajectory for the vehicle to follow based, at least in part, on the route and the at least one perceived object.

15. The computer system of claim 14 , wherein the computer system is provided on an autonomous vehicle.

16. The computer system of claim 14 , wherein the computer system communicates with an autonomous vehicle over a network.