Unlocking vehicle doors with facial recognition

Abstract

The subject disclosure relates to ways to authenticate a rider/user of an autonomous vehicle (AV) using biometric data, such as facial recognition. A process of the disclosed technology can facilitate the automatic unlocking of an AV by performing steps that include: receiving a dispatch request associated with a user identifier (ID), receiving a recognition model that corresponds with the user ID, and receiving an image stream including images of pedestrian faces. In some aspects, the process can further include steps for: providing the images to the recognition model, and determining if a user represented in the images corresponds with the user ID. Systems and machine-readable media are also provided.

What is claimed is: 
     
         1 . A computer-implemented method for automatically unlocking an autonomous vehicle (AV), comprising:
 receiving, at an AV computing system, a first dispatch request, wherein the first dispatch request is associated with a first user identifier (ID);   receiving, at the AV computing system, a first recognition model, wherein the first recognition model corresponds with the first user ID;   receiving, at the AV computing system, an image stream comprising one or more images of pedestrian faces in response to a detected arrival of the AV to a location indicated by the first dispatch request;   providing the one or more images to the first recognition model;   determining, using the first recognition model, if a first user represented in the one or more images corresponds with the first user ID; and   unlocking a door of the AV in response to a match between at least one of the one or more images and the first user ID.   
     
     
         2 . The computer-implemented method of  claim 1 , further comprising:
 updating the first recognition model in response to a match between at least one of the one or more images and the first user ID.   
     
     
         3 . The computer-implemented method of  claim 1 , further comprising:
 receiving a second recognition model corresponding with a second user ID;   providing the one or more images to the second recognition model;   determining, using the second recognition model, if a second user represented by the one or more images corresponds with the second user ID; and   unlocking a door of the AV in response to a match between at least one of the one or more images and the second user ID.   
     
     
         4 . The computer-implemented method of  claim 3 , wherein the second recognition model is a machine-learning model that has been trained using a plurality of facial images of the second user. 
     
     
         5 . The computer-implemented method of  claim 1 , further comprising:
 collecting, at the AV computing system, an image stream in response to reaching first geolocation coordinates; and   wherein the first user ID comprises the first geolocation coordinates associated with the first user.   
     
     
         6 . The computer-implemented method of  claim 1 , wherein the image stream is collected in response to a detected approach of a pedestrian toward the AV. 
     
     
         7 . The computer-implemented method of  claim 1 , wherein the image stream is received from a charge-coupled device (CCD) mounted to the AV. 
     
     
         8 . A system for automatically unlocking an autonomous vehicle (AV), comprising:
 one or more processors;   an image detection device coupled to the one or more processors; and   a computer-readable medium comprising instructions stored therein, which when executed by the processors, cause the processors to perform operations comprising:
 receiving, at an AV computing system, a first dispatch request, wherein the first dispatch request is associated with a first user identifier (ID); 
 receiving, at the AV computing system, a first recognition model, wherein the first recognition model corresponds with the first user ID; 
 receiving, at the AV computing system, an image stream comprising one or more images of pedestrian faces in response to a detected arrival of the AV to a location indicated by the first dispatch request; 
 providing the one or more images to the first recognition model; 
 determining, using the first recognition model, if a first user represented in the one or more images corresponds with the first user ID; and 
 unlocking a door of the AV in response to a match between at least one of the one or more images and the first user ID. 
   
     
     
         9 . The system of  claim 8 , wherein the processors are further configured to perform operations comprising:
 updating the first recognition model in response to a match between at least one of the one or more images and the first user ID.   
     
     
         10 . The system of  claim 8 , wherein the processors are further configured to perform operations comprising:
 receiving a second recognition model corresponding with a second user ID;   providing the one or more images to the second recognition model;   determining, using the second recognition model, if a second user represented by the one or more images corresponds with the second user ID; and   unlocking a door of the AV in response to a match between at least one of the one or more images and the second user ID.   
     
     
         11 . The system of  claim 10 , wherein the second recognition model is a machine-learning model that has been trained using a plurality of facial images of the second user. 
     
     
         12 . The system of  claim 8 , wherein the processors are further configured to perform operations comprising:
 collecting, at the AV computing system, an image stream in response to reaching first geolocation coordinates; and   wherein the first user ID comprises the first geolocation coordinates associated with the first user.   
     
     
         13 . The system of  claim 8 , wherein the image stream is collected in response to a detected approach of a pedestrian toward the AV. 
     
     
         14 . The system of  claim 8 , wherein the image stream is received from a charge-coupled device (CCD) mounted to the AV. 
     
     
         15 . A non-transitory computer-readable storage medium comprising instructions stored therein, which when executed by one or more processors, cause the processors to perform operations comprising:
 receiving, at an AV computing system, a first dispatch request, wherein the first dispatch request is associated with a first user identifier (ID);   receiving, at the AV computing system, a first recognition model, wherein the first recognition model corresponds with the first user ID;   receiving, at the AV computing system, an image stream comprising one or more images of pedestrian faces in response to a detected arrival of the AV to a location indicated by the first dispatch request;   providing the one or more images to the first recognition model;   determining, using the first recognition model, if a first user represented in the one or more images corresponds with the first user ID; and   unlocking a door of the AV in response to a match between at least one of the one or more images and the first user ID.   
     
     
         16 . The non-transitory computer-readable storage medium of  claim 15 , wherein the processors are further configured to perform operations comprising:
 updating the first recognition model in response to a match between at least one of the one or more images and the first user ID.   
     
     
         17 . The non-transitory computer-readable storage medium of  claim 15 , wherein the processors are further configured to perform operations comprising:
 receiving a second recognition model corresponding with a second user ID;   providing the one or more images to the second recognition model;   determining, using the second recognition model, if a second user represented by the one or more images corresponds with the second user ID; and   unlocking a door of an autonomous vehicle (AV) in response to a match between at least one of the one or more images and the second user ID.   
     
     
         18 . The non-transitory computer-readable storage medium of  claim 17 , wherein the second recognition model is a machine-learning model that has been trained using a plurality of facial images of the second user. 
     
     
         19 . The non-transitory computer-readable storage medium of  claim 15 , wherein the processors are further configured to perform operations comprising:
 collecting, at the AV computing system, an image stream in response to reaching first geolocation coordinates; and   wherein the first user ID comprises the first geolocation coordinates associated with the first user.   
     
     
         20 . The non-transitory computer-readable storage medium of  claim 15 , wherein the image stream is collected in response to a detected approach of a pedestrian toward the AV.