X Development LLC

Boston Dynamics, Inc.

Google Inc.

Intrinsic Innovation LLC

Peter Elving Anderson-Sprecher

Geoffrey Lalonde

James Joseph Kuffner

Alex Perkins

Charles DuHadway

Julian Mac Neille Mason

Rohit Ramesh Saboo

Chaitanya Gharpure

Anthony Gerald Francis, Jr

Kevin William Watts

Jr James J Kuffner

A method for detecting boxes includes receiving a plurality of image frame pairs for an area of interest including at least one target box. Each image frame pair includes a monocular image frame and a respective depth image frame. For each image frame pair, the method includes determining corners for a rectangle associated with the at least one target box within the respective monocular image frame. Based on the determined corners, the method includes the following: performing edge detection and determining faces within the respective monocular image frame; and extracting planes corresponding to the at least one target box from the respective depth image frame. The method includes matching the determined faces to the extracted planes and generating a box estimation based on the determined corners, the performed edge detection, and the matched faces of the at least one target box.

Detecting boxes

Methods, apparatus, systems, and computer-readable media are provided for generating a spatio-temporal object inventory based on object observations from mobile robots and determining, based on the spatio-temporal object inventory, monitoring parameters for the mobile robots for one or more future time periods. Some implementations relate to using the spatio-temporal object inventory to determine a quantity of movements of objects that occur in one or more areas of the environment when one or more particular criteria are satisfied&#x2014;and using that determination to determine monitoring parameters that can be utilized to provide commands to one or more of the mobile robots that influence one or more aspects of movements of the mobile robots at future time periods when the one or more particular criteria are also satisfied.

Using object observations of mobile robots to generate a spatio-temporal object inventory, and using the inventory to determine monitoring parameters for the mobile robots

Systems and methods related to roadmaps for robotic devices are provided. A computing device can receive a roadmap representing a plurality of paths through an environment. The computing device can discretize the roadmap to obtain a discrete planning graph having a plurality of states corresponding to discretized segments of the plurality of paths of the roadmap such that states corresponding to adjacent discretized path segments are connected in the discrete planning graph. The computing device can determine a Boolean equation representing at least a portion of the discrete planning graph. The computing device can determine a sequence of states from the plurality of states of the discrete planning graph such that the determined sequence of states satisfies the Boolean equation. The computing device can provide a route through the environment for a robotic device based on the determined sequence of states.

Boolean satisfiability (SAT) reduction for geometry and kinematics agnostic multi-agent planning

A method includes receiving a first time-parameterized path for the first robotic device, and an indication of a second robotic device having a second time-parameterized path that overlaps with the first time-parameterized path at a first location. The method also includes executing, by the first robotic device, a first portion of the first time-parameterized path before reaching the first location, wherein execution of the first portion corresponds to a first rate of progress of the first robotic device along the first time-parameterized path. The first robotic device then receives a communication signal from the second robotic device indicating a second rate of progress of the second robotic device along the second time-parameterized path. The method then includes the first robotic device determining a difference between the first rate of progress and the second rate of progress, and modifying execution of the first time-parameterized path based on the determined difference.

Multi-agent coordination under sparse networking

A method includes receiving first and second coordinated paths for first and second robotic devices. The first coordinated path comprises a dependency edge indicating a first position on the first coordinated path and a second position on the second coordinated path. The method also includes determining a first traversable portion extending to a first stopping position before or at the first position on the first coordinated path. The method also includes providing a first instruction to the first robotic device to traverse the first traversable portion; subsequently determining that the second robotic device has passed the second position on the second coordinated path; determining a second traversable portion of the first coordinated path extending to a second stopping position beyond the first position on the first coordinated path; and providing a second instruction to the first robotic device to traverse the second traversable portion.

Layered multi-agent coordination

A computing device can determine a roadmap having a path for a robotic device in an environment associated with starting and ending poses. The computing device can generate a plurality of trajectories from the starting pose, where each trajectory can include a steering position and a traction velocity directing the robotic device during a planning time interval. For each trajectory of the plurality of trajectories, the computing device can determine a score for the trajectory indicative of advancement from the starting pose toward the ending pose after simulating the steering position and the traction velocity for the planning time interval. The computing device can select, and then store, a nominal trajectory from among the scored plurality of trajectories. The computing device can receive a first request to provide a route though the environment and can send a first response that includes the stored nominal trajectory.

Pre-computation of kinematically feasible roadmaps

Methods, apparatus, systems, and computer-readable media are provided for using sensor-based observations from multiple agents (e.g., mobile robots and/or fixed sensors) in an environment to estimate the pose of an object in the environment at a target time and to estimate an uncertainty measure for that pose. Various implementations generate a multigraph based on a group of observations from multiple agents, where the multigraph includes a reference frame node, object nodes, and a plurality edges connecting the nodes. In some implementations, a composite pose and composite uncertainty measure are generated for each of a plurality of simple paths along the edges of the multigraph that connect the reference frame node to a given object node&#x2014;and a pose and uncertainty measure for an object identifier associated with the given object node is generated based on the composite poses and the composite uncertainty measures.

Using sensor-based observations of agents in an environment to estimate the pose of an object in the environment and to estimate an uncertainty measure for the pose

Growing community of inventors

Mountain View, CA, United States of America

Peter Elving Anderson-Sprecher