Recognizing assigned passengers for autonomous vehicles
Aspects of the disclosure provide systems and methods for recognizing an assigned passenger. For instance, dispatching instructions to pick up a passenger at a pickup location are received. The instructions include authentication information for authenticating a client computing device associated with the passenger. A vehicle is maneuvered in an autonomous driving mode towards the pickup location. The client device is then authenticated. After authentication, a set of pedestrians within a predetermined distance of the vehicle are identified from sensor information generated by a sensor of the vehicle and location information is received over a period of time from the client device. The received location information is used to estimate a velocity of the passenger. This estimated velocity is used to identify a subset of set of pedestrians that is likely to be the passenger. The vehicle is stopped to allow the passenger to enter the vehicle based on the subset.
1. A method of recognizing an assigned passenger, the method comprising:
receiving, by one or more processors of a vehicle, dispatching instructions to pick up the assigned passenger at a pickup location;
maneuvering, by the one or more processors, the vehicle towards the pickup location in an autonomous driving mode;
identifying, by the one or more processors, from sensor information generated by a sensor of the vehicle, a set of pedestrians that are within a predetermined distance of the pickup location;
receiving, by the one or more processors, orientation information for a client computing device associated with the assigned passenger;
receiving, by the one or more processors, orientation information indicative of the orientations of pedestrians of the set of pedestrians;
comparing, by the one or more processors, the orientation information for the client computing device to the orientation information for the set of pedestrians; and
based on the comparing, identifying, by the one or more processors, one of the pedestrians of the set of pedestrians as the assigned passenger.
2. The method of claim 1 , further comprising using the sensor information to detect a gaze direction for each pedestrian of the set of pedestrians, and wherein identifying the one of the pedestrians is further based on the gaze direction for each pedestrian.
3. The method of claim 2 , wherein identifying the one of the pedestrians is further based on whether the gaze direction for each pedestrian indicates that the pedestrian is looking at the vehicle.
4. The method of claim 1 , further comprising using the sensor information to determine a number of other pedestrians corresponding to pedestrians within a predetermined distance of each pedestrian of the set of pedestrians, and wherein identifying the one of the pedestrians is further based on the determined number of other pedestrians within the predetermined distance of each pedestrian.
5. The method of claim 4 , wherein the dispatching instructions further identify a number of passengers, and wherein identifying the one of the pedestrians is further based on the identified number of passengers.
6. The method of claim 1 , wherein the set of pedestrians is updated as additional orientation information is received from the client computing device.
7. The method of claim 1 , further comprising, based on the comparing, stopping, by the one or more processors, the vehicle to allow the assigned passenger to enter the vehicle.
8. The method of claim 7 , wherein stopping the vehicle includes stopping the vehicle closer to a pedestrian of the set than to the pickup location.
9. The method of claim 7 , wherein the dispatching instructions further include a pickup location, and stopping the vehicle includes stopping the vehicle before the vehicle reaches the pickup location.
10. The method of claim 1 , further comprising:
using the sensor information to identify a characteristic that is different between two or more pedestrians of the set of pedestrians;
sending a request to the client device, the request including a question regarding the characteristic; and
receiving a response from the client computing device, and wherein identifying the one of the pedestrians is further based on the response.
11. The method of claim 1 , wherein the comparing includes inputting the orientation information for the client computing device and the orientation information for the set of pedestrians into a model in order to identify a likelihood that each pedestrian of the set of pedestrians is the assigned passenger, and wherein identifying the one of the pedestrians is further based on the likelihood that each pedestrian of the set of pedestrians is the passenger.
12. The method of claim 1 , wherein the orientation information of the client computing device is generated by a sensor of the client computing device.
13. The method of claim 1 , wherein the sensor is at least one of an accelerometer or a gyroscope.
14. The method of claim 1 , wherein the orientation information for the client computing device includes a heading of the client computing device.
15. The method of claim 1 , wherein the orientation information for the set of pedestrians includes a heading for each of each pedestrian of the set of pedestrians.
16. The method of claim 1 , wherein the set of pedestrians are identified based on a predetermined distance of the vehicle.
17. The method of claim 1 , wherein the set of pedestrians includes at least two pedestrians.
18. The method of claim 17 , wherein one of the at least two pedestrians is the assigned passenger and another of the at least two pedestrians is not the assigned passenger.
19. The method of claim 1 , further comprising, filtering at least one pedestrian from the set of pedestrians which is unlikely to be waiting for the vehicle.