Identifying a route for an autonomous vehicle between an origin and destination location
Described herein are technologies relating to computing a likelihood of an operation-influencing event with respect to an autonomous vehicle at a geographic location. The likelihood of the operation-influencing event is computed based upon a prediction of a value that indicates whether, through a causal process, the operation-influencing event is expected to occur. The causal process is identified by means of a model, which relates spatiotemporal factors and the operation-influencing events.
1. A computing system that is in network communication with a computing device, the computing system comprising:
a processor; and
memory storing instructions that, when executed by the processor, cause the processor to perform acts comprising:
obtaining an origin location of an autonomous vehicle (AV) and a destination location of the AV from the computing device, where at least one of the origin location or the destination location is within a city;
obtaining a first value for a spatiotemporal factor at a first location within the city;
computing a second value for the spatiotemporal factor at a second location within the city, where the second location is along a roadway within a candidate route between the origin location and the destination location, where the candidate route is amongst several candidate routes between the origin location and the destination location, and further where the second value is computed based upon:
the first value for the spatiotemporal factor at the first location; and
a distance between the first location and the second location;
computing a likelihood of occurrence of a predefined event at the second location based upon the computed second value, where the predefined event relates to travel of the vehicle between the origin location and the destination location;
selecting the candidate route from amongst the several candidate routes, where the candidate route is selected based upon the computed likelihood of occurrence of the predefined event at the second location; and
transmitting the candidate route between the origin location and the destination location to the computing device, where the computing device controls at least one of a propulsion system, a braking system, or a steering system of an autonomous vehicle (AV) to cause the AV to travel over the candidate route.
2. The computing system of claim 1 , wherein the computing device is onboard the AV.
3. The computing system of claim 2 , wherein the predefined event is a human takeover of the AV at the second location.
4. The computing system of claim 2 , wherein the predefined event is a sudden stop of the AV.
5. The computing system of claim 2 , wherein the first value for the spatiotemporal factor is at least one of:
a number of pedestrians observed by a second AV when the second AV was at the first location;
an indication as to whether the second AV observed steam at the first location;
an indication as to whether the second AV observed an emergency vehicle at the first location; or
an indication as to whether the second AV observed a double parked vehicle at the first location.
6. The computing system of claim 1 , wherein the likelihood of occurrence of the predefined event at the second location is computed by a mixed-effects model.
7. The computing system of claim 1 , wherein the first location is a first road intersection within the city and the second location is a second road intersection within the city.
8. The computing system of claim 1 , wherein the distance between the first location and the second location is a Euclidean distance between the first location and the second location.
9. The computing system of claim 1 , wherein the distance between the first location and the second location is a distance along a shortest path over roadways between the first location and the second location.
10. The computing system of claim 1 , the acts further comprising:
computing an estimated travel time between the origin location and the destination location over the candidate route, wherein the candidate route is selected based further upon the estimated travel time.
11. A method performed by a computing system that is in communication with a computing device, the method comprising:
obtaining an origin location of a vehicle and a destination location of the vehicle, where the vehicle is to travel over roadways from the origin location to the destination location, and further where at least one of the origin location or the destination location is within a city;
obtaining a first value for a spatiotemporal factor at a first location within the city;
computing a second value for the spatiotemporal factor at a second location within the city, where the second location is along a roadway within a candidate route from amongst several candidate routes between the origin location and the destination location, where the second value for the spatiotemporal factor is computed based upon:
the first value for the spatiotemporal factor at the first location; and
a distance between the first location and the second location;
computing a likelihood value that is indicative of a likelihood of a predefined event occurring at the second location at a time that the vehicle is estimated to be at the second location when travelling the candidate route between the origin location and the destination location, where the likelihood value is computed based upon the second value for the spatiotemporal factor at the second location;
selecting the candidate route from amongst the several candidate routes based upon the likelihood value; and
transmitting the candidate route between the origin location and the destination location to the computing device based upon the candidate route being selected, where the computing device controls at least one of a propulsion system, a braking system, or a steering system of the vehicle based upon the candidate route between the origin location and the destination location.
12. The method of claim 11 , further comprising:
obtaining a first value for a second spatiotemporal factor at the first location;
computing a second value for the second spatiotemporal factor at the second location, where the second value for the second spatiotemporal factor is computed based upon:
the first value for the second spatiotemporal factor at the first location; and
the distance between the first location and the second location, wherein the likelihood value is computed based upon the second value for the second spatiotemporal factor.
13. The method of claim 11 , further comprising:
computing an estimated travel time for the candidate route based upon the origin location and the destination location, wherein the candidate route is selected based upon the estimated travel time.
14. The method of claim 11 , where the vehicle is an autonomous vehicle (AV), and further wherein the predefined event relates to travel of the AV at the second location.
15. The method of claim 11 , where the first location is a first road segment in a road network of the city, and further where the second location is a second road segment in the road network of the city.
16. The method of claim 11 , wherein a mixed-effects model is used to compute the likelihood value.
17. The method of claim 11 , wherein the first value for the spatiotemporal factor is a number of pedestrians observed at the first location by an autonomous vehicle (AV) travelling at the first location.
18. The method of claim 11 , wherein the predefined event is a sudden stop of the vehicle.
19. An autonomous vehicle (AV) comprising a computer-readable storage medium that, when executed by a processor of the AV, causes the AV to perform acts comprising:
obtaining an origin location of the AV and a destination location of the AV, where the AV is to travel over roadways from the origin location to the destination location, and further where at least one of the origin location or the destination location is within a city;
obtaining a first value for a spatiotemporal factor at a first location within the city;
computing a second value for the spatiotemporal factor at a second location within the city, where the second location is along a roadway within a candidate route from amongst several candidate routes between the origin location and the destination location, where the second value for the spatiotemporal factor is computed based upon:
the first value for the spatiotemporal factor at the first location; and
a distance between the first location and the second location;
computing a likelihood value that is indicative of a likelihood of a predefined event occurring at the second location at a time that the AV is estimated to be at the second location when travelling the candidate route between the origin location and the destination location, where the likelihood value is computed based upon the second value for the spatiotemporal factor at the second location;
selecting the candidate route from amongst the several candidate routes based upon the likelihood value; and
controlling at least one of a propulsion system, a braking system, or a steering system of the AV to cause the AV to travel over the candidate route between the origin location and the destination location.
20. The AV of claim 19 , where the first location is a first road segment in a road network of the city, and further where the second location is a second road segment in the road network of the city.