Skydio, Inc.

Ryan David Kennedy

Peter Benjamin Henry

Abraham Galton Bachrach

Hayk Martirosyan

Adam Parker Bry

Jack Louis Zhu

Harrison Zheng

Matthew Joseph Donahoe

Brian Richman

Himel Mondal

Quentin Allen Wah Yen Delepine

Kristen Marie Holtz

Gareth Benoit Cross

Samuel Shenghung Wang

Teodor Tomic

Josiah Timothy VanderMey

Dominic William Pattison

Vladimir Nekrasov

Bradley Matthew Solliday

Nathan Leo Bucki

Yixiao Sun

Aaron Christopher Miller

Clement Godard

In some examples, a computing device receives, from an unmanned aerial vehicle (UAV), a first image from a first camera on the UAV and a plurality of second images from a plurality of second cameras on the UAV. The plurality of second cameras may be positioned on the UAV for providing a plurality of different fields of view in a plurality of different directions around the UAV. Further, the first camera has a longer focal length than the second cameras. The computing device presents, on a display, a composite image including at least a portion of the first image within a merged image generated from the plurality of second images. The presented composite image enables a user to at least one of: zoom out from the at least one first image to the merged image, or zoom in from the merged image to the at least one first image.

Unmanned aerial vehicle with virtual un-zoomed imaging

A computer of an unmanned aerial vehicle (UAV) accesses, from a memory unit, a problem definition comprising cost functions associated with travel of the UAV. The computer causes movement of the UAV based on the cost functions. The computer adjusts one or more of the cost functions during a flight of the UAV. The computer causes further movement of the UAV based on the adjusted one or more of the cost functions.

Unmanned aerial vehicle operated based on cost functions

A computer accesses an input element storage and an output element storage. The computer accesses a symbolic expression for output element storage as a function of the input element storage. The computer computes, using a symbolic computation engine of the computer, a symbolic expression for the tangent space Jacobian of the output element storage with respect to an input tangent space. The computer outputs the computed expression.

Computer-implemented symbolic differentiation using first-order retraction

An autonomous vehicle that is equipped with image capture devices can use information gathered from the image capture devices to plan a future three-dimensional (3D) trajectory through a physical environment. To this end, a technique is described for image-space based motion planning. In an embodiment, a planned 3D trajectory is projected into an image-space of an image captured by the autonomous vehicle. The planned 3D trajectory is then optimized according to a cost function derived from information (e.g., depth estimates) in the captured image. The cost function associates higher cost values with identified regions of the captured image that are associated with areas of the physical environment into which travel is risky or otherwise undesirable. The autonomous vehicle is thereby encouraged to avoid these areas while satisfying other motion planning objectives.

Image space motion planning of an autonomous vehicle

A computer accesses a first symbolic expression for an output matrix as a function of an input matrix at a computing device comprising processing circuitry and memory. The computer computes a first Jacobian of the input matrix with respect to an input tangent space. The computer computes a second Jacobian of the output matrix with respect to the input matrix. The computer computes a third Jacobian of an output tangent space with respect to the input matrix. The computer applies symbolic matrix multiplication to the first Jacobian, the second Jacobian, and the third Jacobian to obtain a second symbolic expression for the output tangent space with respect to the input tangent space. The computer provides a representation of the second symbolic expression, the second symbolic expression representing a computed tangent-space Jacobian.

Computer-implemented computation of tangent-space jacobian

Autonomous aerial navigation in low-light and no-light conditions includes using night mode obstacle avoidance intelligence and mechanisms for vision-based unmanned aerial vehicle (UAV) navigation to enable autonomous flight operations of a UAV in low-light and no-light environments using infrared data.

Autonomous aerial navigation in low-light and no-light conditions

Autonomous aerial navigation in low-light and no-light conditions includes using night mode obstacle avoidance intelligence, training, and mechanisms for vision-based unmanned aerial vehicle (UAV) navigation to enable autonomous flight operations of a UAV in low-light and no-light environments using infrared data.

In some examples, one or more processors of an unmanned aerial vehicle (UAV), control a propulsion mechanism of the UAV to cause the UAV to navigate to a plurality of positions in relation to a scan target. Using one or more image sensors of the UAV, a first image of the scan target is captured from a first position of the plurality of positions, and a second image of the scan target is captured from a second position of the plurality of positions. A disparity is determined between the first image captured at the first position and the second image captured at the second position. A three-dimensional model corresponding to the scan target is determined based in part on the disparity determined between the first image and the second image.

Determining a three-dimensional model of a scan target

In some examples, one or more processors of an aerial vehicle access a scan plan including a sequence of poses for the aerial vehicle to assume to capture, using the one or more image sensors, images of a scan target. A next pose of the scan plan is checked for obstructions, and based at least on detection of an obstruction, the one or more processors determine whether a backup pose is available for capturing an image of the targeted point orthogonally along a normal of the targeted point. Responsive to determining that the backup pose is unavailable for capturing an image of the targeted point orthogonally along the normal of the targeted point, image capture of the targeted point is performed at an oblique angle to the normal of the targeted point.

Unmanned aerial vehicle adaptable to obstructions

In some examples, an image of a scan target is presented in a user interface on a display associated with a computing device. The user interface receives at least one user input indicating at least one point in a perimeter or edge of a volume for encompassing the scan target presented in the image of the scan target. A graphical representation of the volume in relation to the image of the scan target is generated in the user interface. Information for defining a location of at least a portion of the volume in three-dimensional space is sent to an unmanned aerial vehicle (UAV) to cause, at least in part, the UAV to scan at least a portion of the scan target corresponding to the volume.

Indicating a scan target for an unmanned aerial vehicle

In some examples, an unmanned aerial vehicle (UAV) may include one or more processors configured to capture, with one or more image sensors, and while the UAV is in flight, a plurality of images of a target. The one or more processors may compare a first image of the plurality of images with a second image of the plurality of images to determine a difference between a current frame of reference position for the UAV and an estimate of an actual frame of reference position for the UAV. In addition, the one or more processors may determine, based at least on the difference, and while the UAV is in flight, an update to a three-dimensional model of the target.

Performing 3D reconstruction via an unmanned aerial vehicle

In some examples, an unmanned aerial vehicle (UAV) may identify a scan target. The UAV may navigate to two or more positions in relation to the scan target. The UAV may capture, using one or more image sensors of the UAV, two or more images of the scan target from different respective positions in relation to the scan target. For instance, the two or more respective positions may be selected by controlling a spacing between the two or more respective positions to enable determination of parallax disparity between a first image captured at a first position and a second image captured at a second position of the two or more positions. The UAV may determine a three-dimensional model corresponding to the scan target based in part on the determined parallax disparity of the two or more images including the first image and the second image.

Growing community of inventors

San Francisco, CA, United States of America

Ryan David Kennedy