US 12,138,812 B2
Systems and methods for laser and imaging odometry for autonomous robots
Girish Bathala, San Diego, CA (US); Sahil Dhayalkar, San Diego, CA (US); Kirill Pirozhenko, San Diego, CA (US); and Oleg Sinyavskiy, San Diego, CA (US)
Assigned to Brain Corporation, San Diego, CA (US)
Filed by Brain Corporation, San Diego, CA (US)
Filed on Jul. 27, 2021, as Appl. No. 17/386,001.
Application 17/386,001 is a continuation of application No. PCT/US2020/016070, filed on Jan. 31, 2020.
Claims priority of provisional application 62/799,142, filed on Jan. 31, 2019.
Prior Publication US 2021/0354302 A1, Nov. 18, 2021
Int. Cl. B25J 9/16 (2006.01); B25J 13/08 (2006.01); G06T 7/20 (2017.01); G06T 7/579 (2017.01); G06T 7/70 (2017.01)
CPC B25J 9/1697 (2013.01) [B25J 9/1664 (2013.01); B25J 13/089 (2013.01); G06T 7/20 (2013.01); G06T 7/579 (2017.01); G06T 7/70 (2017.01); G06T 2207/20221 (2013.01); G06T 2207/30241 (2013.01)] 16 Claims
OG exemplary drawing
 
1. A method for determining a pose of a sensor of a robot, comprising:
generating, via a controller configured to execute computer readable instructions, a composite image including overlaying of a first image and a second image captured by the sensor of the robot at different time intervals, the first image including a plurality of pixels representing an object at a first location, and the second image including a plurality of pixels representing the object at a different second location;
determining, via the controller, a first motion of the robot based on a discrepancy between the first and second images by evaluating at least one depth disparity measurement and at least one spatial disparity measurement of the object represented in the first and second images;
determining, via the controller, a second motion of the robot based on odometry data from one or more proprioceptive sensor of the robot;
determining, via the controller, a most probable pose of the sensor based at least in part on a plurality of possible poses of the sensor and a continuous distribution function of the plurality of possible poses by applying a maximum a posteriori estimation, wherein the continuous distribution function of the plurality of possible poses comprises a set of priors for the maximum a posteriori estimation, wherein each of the plurality of possible poses is determined based on a respective discrepancy of a plurality of discrepancies between the first motion and the second motion, and wherein the continuous distribution function of the plurality of possible poses is measured over a period of time beginning prior to the first motion and the second motion of the robot; and
localizing, via the controller, the object on a computer readable map based at least in part on the most probable pose of the sensor.