1.2Posture at first contact
A collision solver advances the object from its recent motion samples until a foot collider reaches the ground surface. The renderer draws the joint angles at that contact event as a wireframe.
Generated automatically. Not reviewed by a person. Not legal advice or an opinion.
Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.
A collision solver advances the object from its recent motion samples until a foot collider reaches the ground surface. The renderer draws the joint angles at that contact event as a wireframe.
The predictor copies the current joint angles and velocities from the physics engine and holds the sampled controller input constant. It advances that copy through six simulation steps of 1/60 second and sends the final joint angles to the preview renderer.
The system estimates the object's velocity from recent center-of-mass positions and projects a trajectory under gravity. It draws a ring at the first intersection with the ground surface.
The renderer draws both postures with the same mesh, materials, lighting, camera settings, and opacity. The two views carry fixed labels that identify the current posture and the prediction time.
The system extracts torso roll from the predicted joint angles and activates the vibration motor on the side of the lean. Pulse strength rises with the roll magnitude up to 30 degrees, and a two-degree dead band keeps small changes from triggering pulses.
Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.
A warehouse training simulator records the position, heading and mast tilt of a trainee-operated virtual forklift on each simulation tick. The simulator estimates travel speed, turn rate and mast tilt rate from the latest samples and integrates those rates over the next half second. The display draws the predicted chassis and mast as wireframes over the solid current forklift in the same aisle view.
Body-mounted inertial measurement units supply the motion history for a player-controlled avatar. The processor applies each sensor's mounting calibration, converts the sensor orientations to parent-relative joint orientations and buffers the results with timestamps. It estimates each joint's angular velocity from successive orientations and integrates it over the prediction interval to pose a translucent copy beside the solid current avatar.
The game engine retains 24 joint pose samples at 120 samples per second for the player-controlled character. A predictor estimates linear and angular velocities from the buffer and projects the character's root and joints to 50, 150 and 300 milliseconds ahead. The renderer draws the current character as a solid mesh and each forecast as an outline with a distinct line pattern and a time label.
A game replay system stores timestamped skeletal poses of the player-controlled character and exposes a frame cursor. When the cursor moves, the predictor estimates joint velocities from the selected frame and preceding recorded frames, then advances the pose by 200 milliseconds. The replay viewport draws the selected recorded pose as a solid mesh and overlays the projected pose as a wireframe.
The game engine extrapolates relative joint orientations from recent pose samples at candidate horizons of 80, 160 and 240 milliseconds. As the corresponding simulation frames arrive, it compares each stored projection with the resulting pose and updates a rolling mean angular error for that horizon. The renderer shows a translucent pose at the longest horizon whose mean error is below 8 degrees, or at the shortest horizon when all three errors reach or exceed the threshold.
The engine estimates foot trajectories and pelvis motion from recent skeletal poses of the player-controlled character and projects them 180 milliseconds forward. It sweeps a foot-shaped collision volume along each descending foot trajectory against the terrain mesh to find a contact point and surface normal. For each detected contact, a two-bone inverse kinematics solver adjusts the predicted hip and knee to place the foot at that point, and the renderer overlays the corrected pose as a translucent mesh.
A hardwired arithmetic pipeline reads recent joint samples and writes projected joint angles to dual-port memory. A separate raster circuit reads those angles at each frame boundary and draws the preview in a second color.
A processor runs the posture predictor once per simulation tick and writes the predicted joint transforms into one of two buffers. It submits a graphics command that reads the completed buffer while the predictor writes to the other buffer.
The predictor stores eight joint-angle samples per joint at 60 samples per second and fits a straight line to each set of samples after unwrapping the angles. It projects each line 120 milliseconds beyond the latest sample, clamps the angles to the joint limits, and assembles the predicted posture.
The renderer draws the current posture as an opaque mesh and skins a second instance of that mesh with the predicted joint transforms. It draws the second mesh at 25 percent opacity with depth testing enabled and depth writes disabled.
The renderer draws the current object with its shaded surface mesh. It projects the predicted joint positions into screen coordinates and connects adjacent joints with dashed lines using six-pixel dashes and four-pixel gaps.
Passages reproduced word for word from two earlier-published U.S. patent documents, placed beside the language of claim 1. No characterization is offered. Each passage is marked A or B for the document it comes from; ¶ numbers are the paragraph numbers in the USPTO’s published full text of that document.
The example computer system 1200 includes a processor 1202 (e.g., a central processing unit (CPU) a graphics processing unit (GPU) or both), a main memory 1204 and a static memory 1206, which communicate with each other via a bus 1208.A · US 9,883,090 B2 · ¶ 64
The posture prediction unit 300 receives the posture command signal and filters a posture prediction through angular velocity limitation and the time-delayed filtering, thereby simulating/predicting a response from the autopilot control loop 200 and outputting a posture prediction signal (S300).B · US 10,082,800 B2 · ¶ 19
Once the picture is taken, it is saved and the camera application (or item listing application) can advance to the next wireframe, depicted as 604 in FIG. 6C. Here, the same live image as was taken in FIG. 6B is still present, but a new wireframe 604 is overlayed, showing the user the orientation of the next photo to take.A · US 9,883,090 B2 · ¶ 58
An empty cell means only that no passage was reproduced there.
The cameras on these mobile devices can come in handy when photographing the item or items to be sold, and the ability to have the camera integrated with the same device where the user is entering information (through, for example, a stand-alone application installed on the mobile device), makes the experience more seemless for users.A · US 9,883,090 B2 · ¶ 3
Individual sellers, however, often photograph these items in wildly variable (and often ineffective) ways.A · US 9,883,090 B2 · ¶ 3
However, the mission equipment has problems that it is very expensive, is slow to respond to a sudden maneuver of the aerial system and shows stabilization performance varied depending on flight characteristics.B · US 10,082,800 B2 · ¶ 3
(f) by the differentiation filter 500, removing noise from the differentiated angular velocity command signal; and (g) by the mission equipment 600, receiving the angular velocity command signal and stably doing a mission.B · US 10,082,800 B2 · ¶ 6
patent.fyi, US 2026/0355632 A1, claim 1, first published 2026-11-19T13:37:15-05:00, https://patent.fyi/US20260355632/, SHA-256 68d2c3d426126d25018d387e448cac8823be5a3c1f1b00ed66677a701d20a147