Robot Docking Station Identification Surface

Abstract

A docking station is provided that includes at least one component configured to couple to a robot and an identification surface. The identification surface includes a first curvature that varies at a first substantially constant rate of change along a first dimension the identification includes a second curvature that varies at a second substantially constant rate of change along a second dimension. The second dimension is orthogonal to the first dimension. The identification surface includes a third curvature that varies at a third substantially constant rate of change along a third dimension. The third dimension is orthogonal to the first dimension and the second dimension.

What is claimed is: 
     
         1 . A method comprising:
 receiving, from one or more sensors of a robot, sensor data indicative of an environment of the robot;   determining, based on the sensor data, a local curvature of a surface along each of a first dimension, a second dimension, and a third dimension;   determining that the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension matches a representation of a known identification surface of a docking station for the robot; and   controlling the robot to navigate relative to the docking station based on determining that the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension matches the representation of the known identification surface of the docking station.   
     
     
         2 . The method of  claim 1 , wherein determining the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension comprises:
 determining, based on the sensor data, a depth map representative of the environment; and   identifying, using the depth map, the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension within the environment.   
     
     
         3 . The method of  claim 2 , further comprising:
 transforming a portion of the depth map to align with the representation of the known identification surface of the docking station for the robot.   
     
     
         4 . The method of  claim 3 , further comprising determining relative distances of a plurality of points of the transformed portion of the depth map to a corresponding plurality of points of the representation of the known identification surface, wherein determining that the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension matches the representation of the known identification surface comprises determining that the relative distances of the plurality of points of the transformed portion of the depth map to the corresponding plurality of points of the representation of the known identification surface are each less than a threshold distance. 
     
     
         5 . The method of  claim 2 , wherein controlling the robot to navigate relative to the docking station comprises:
 determining a pose of the robot relative to the docking station based on which portion of the representation of the known identification surface matches the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension; and   navigating toward the docking station based on the pose of the robot relative to the docking station.   
     
     
         6 . The method of  claim 5 , wherein the depth map is a first depth map representative of the environment at a first time, the method further comprising:
 determining, based on the sensor data, a second depth map representative of the environment at a second time;   identifying, using the second depth map, the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension within the environment at the second time;   determining that a portion of the second depth map corresponding to the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension at the second time matches a second portion of the representation of the known identification surface;   determining a second pose of the robot relative to the docking station based on which portion of the representation of the known identification surface matches the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension at the second time; and   navigating toward the docking station based on the second pose of the robot relative to the docking station.   
     
     
         7 . The method of  claim 1 , further comprising:
 prior to determining the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension, determining that a charge level of a battery of the robot has dropped below a threshold charge level; and   responsive to determining that the charge level of the battery of the robot has dropped below the threshold charge level, determining a region of interest within the environment associated with the docking station, wherein the sensor data is representative of the region of interest within the environment.   
     
     
         8 . The method of  claim 7 , wherein determining the region of interest comprises:
 determining a pose of the robot within the environment relative to a mapped position of the docking station within the environment; and   determining the region of interest based on the pose of the robot within the environment relative to the mapped position of the docking station.   
     
     
         9 . The method of  claim 7 , wherein determining the region of interest comprises:
 applying a trained machine learning model to image data to determine an initial estimate of a location of the docking station.   
     
     
         10 . The method of  claim 9 , further comprising:
 determining that the initial estimate of the location of the docking station is approximately correct; and   responsive to determining that the initial estimate of the location of the docking station is approximately correct, providing an indication that the trained machine learning model successfully identified the region of interest.   
     
     
         11 . The method of  claim 1 , wherein the docking station comprises a charging station configured to charge a battery of the robot. 
     
     
         12 . The method of  claim 11 , wherein the charging station comprises a positive electric coupling component and a negative electric coupling component corresponding to a set of electric couplings on the robot. 
     
     
         13 . The method of  claim 12 , wherein the positive electric coupling component and the negative electric coupling component couple with the set of electric couplings on the robot to provide electric power to the robot. 
     
     
         14 . The method of  claim 1 , wherein the docking station further comprises a housing, and wherein the known identification surface is disposed on a top portion of the housing. 
     
     
         15 . The method of  claim 14 , wherein the known identification surface is convex to a surrounding environment, and wherein the known identification surface is concave to the housing. 
     
     
         16 . The method of  claim 1 , wherein the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension are all different. 
     
     
         17 . A robot, comprising:
 one or more sensors;   a computing having one or more processors;   a non-transitory computer readable medium; and   program instructions stored on the non-transitory computer readable medium and executable by the one or more processors to:
 receive, from the one or more sensors, sensor data indicative of an environment of the robot; 
 determine, based on the sensor data, a local curvature of a surface along each of a first dimension, a second dimension, and a third dimension; 
 determine that the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension matches a representation of a known identification surface of a docking station for the robot; and 
 control the robot to navigate relative to the docking station based on determining that the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension matches the representation of the known identification surface of the docking station. 
   
     
     
         18 . The robot of  claim 17 , wherein the program instructions stored on the non-transitory computer readable medium and executable by the one or more processors includes to:
 determine, based on the sensor data, a depth map representative of the environment; and   identify, using the depth map, the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension within the environment.   
     
     
         19 . The robot of  claim 18 , wherein the program instructions stored on the non-transitory computer readable medium and executable by the one or more processors further includes to:
 transform a portion of the depth map to align with the representation of the known identification surface of the docking station for the robot.   
     
     
         20 . A non-transitory computer-readable medium having stored therein instructions executable by one or more processors to cause a computing system to perform functions comprising:
 receiving, from one or more sensors of a robot, sensor data indicative of an environment of the robot;   determining, based on the sensor data, a local curvature of a surface along each of a first dimension, a second dimension, and a third dimension;   determining that the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension matches a representation of a known identification surface of a docking station for the robot; and   controlling the robot to navigate relative to the docking station based on determining that the local curvature of the surface along each of the first dimension, the second dimension, and the third dimension matches the representation of the known identification surface of the docking station.