US 12,222,723 B2
Directed exploration for navigation in dynamic environments
Brian Masao Yamauchi, Boston, MA (US)
Assigned to Boston Dynamics, Inc., Waltham, MA (US)
Filed by Boston Dynamics, Inc., Waltham, MA (US)
Filed on Jun. 2, 2022, as Appl. No. 17/805,135.
Claims priority of provisional application 63/202,289, filed on Jun. 4, 2021.
Prior Publication US 2022/0390950 A1, Dec. 8, 2022
Int. Cl. G05D 1/00 (2024.01); G01C 21/00 (2006.01)
CPC G05D 1/0214 (2013.01) [G01C 21/3826 (2020.08); G05D 1/0274 (2013.01)] 26 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
receiving, by data processing hardware of a legged robot, a navigation route of a topological map for the legged robot, the navigation route comprising a set of waypoints and a first waypoint edge connecting a first waypoint of the set of waypoints and a second waypoint of the set of waypoints;
identifying, by the data processing hardware, for a foot of the legged robot, a set of footfall placement locations based on the navigation route;
obtaining, by the data processing hardware, from a sensor of the legged robot, sensor data;
generating, by the data processing hardware, a local obstacle map based on the sensor data, wherein the local obstacle map is different from the topological map;
determining, by the data processing hardware, that an action to touch down the foot at a footfall placement location of the set of footfall placement locations is obstructed by an obstacle, wherein the footfall placement location corresponds to the first waypoint edge;
determining, by the data processing hardware, that the first waypoint edge of the navigation route is obstructed based on determining that the action is obstructed;
generating, by the data processing hardware, using the local obstacle map, a second waypoint edge connecting a position of the legged robot and a third waypoint based on determining that the first waypoint edge of the navigation route is obstructed, wherein the second waypoint edge generated using the local obstacle map avoids the obstacle and comprises a deviation from the navigation route; and
instructing, by the data processing hardware, the legged robot to navigate to the third waypoint via the second waypoint edge generated using the local obstacle map and back to the navigation route.