Feature videos¶
Each clip is a real run, seen through the headless viewer: tools/videos/make_videos.py starts the
simulator, sends the commands and puts each under the picture as it is sent. Remake them with
python3 tools/videos/make_videos.py [CLIP...] after sourcing the workspace (real time on an Intel
Iris Xe, about 12 minutes for all).
Video |
Shows |
Command in the clip |
|---|---|---|
The sea: waves, sun glitter, reflections, foam round a floating crate |
|
|
The same sea getting rough as the wind rises |
|
|
Chrome to rubber, in sun and shadow, with a pane of glass |
|
|
The town: lights change, people press a button and cross, cars stop and go |
|
|
Street level at a lone crossing with its signal |
|
|
Summer |
|
|
Fall |
|
|
Winter |
|
|
The rainy season |
|
|
A robot in the sample warehouse |
|
|
Speed in a slice along an Ahmed body, moving, over pressure seen from above |
|
|
Speed along the Lexus RX 450h’s body mesh, moving, over pressure from above. Made by hand, see cfd.md §6.2 |
|
|
The mean flow as a world: streamlines, a slice, surface pressure, colour scale |
|
|
The quadcopter coming down to half a metre: the outwash ring spreads wide |
|
|
Live air: a lattice Boltzmann box round the drone (4 cm cells, on the GPU), hovering then down to half a metre, where the jets meet under it and rise |
|
|
Beside a 12 m building and over its roof: the wake hugs the wall, spreads on the roof; the rotors by the wall make less thrust |
|
|
The air under a hovering quadcopter: particles in 3D and its speed through the drone, then leaning in wind |
|
|
A pendulum under Earth, Moon and Jupiter gravity, changed while it swings |
|
|
A hovering quadcopter leaning into 6 and 14 m/s of wind, and calm again |
|
|
A time-of-flight sensor from a shared library, its range on the scope |
|
|
Yaw rate on time and 0.4 s late, lidar returns as they arrive |
|
|
A wooden crate bobs at its draft, a steel one sinks |
|
|
Retroreflector, white paint, asphalt and black rubber walls to a lidar |
|
|
No returns from a pond at a slant, also not from the ground under it |
|
|
A spinning lidar in a square room, distortion on and off side by side |
|
|
An arm by position and by velocity per joint |
|
|
A DC motor at 6 V, 12 V and 0 V: speed and current |
|
|
Robots and a person added to and removed from a running simulation |
|
|
One file starting a whole run |
|
|
The sedan’s 4D imaging radar in the town: its detections in the 3D view and in the Sensors workspace’s radar view, white while it stands, blue (closing) once it drives |
|
|
The AMR’s gyro biased by 0.5 rad/s, then stuck while the robot turns the other way, against the true yaw rate |
|
|
The same room through a channel table, a flash lidar and a Livox-style rosette, side by side |
|
|
A 90° pinhole and a 220° equidistant fisheye on the AMR in the warehouse, driving in step |
|
|
A camera and an event camera with the same pixels: the events are the edges that move |
|
|
The safety AMR driven up to a rack and back: its scanner’s nearest return and warning field, and a wheel’s torque |
|
|
The robot builder: a box and a cylinder on one box, a revolute joint posed by its slider, a lidar and a rangefinder from the catalog, an L-shaped mesh part colliding as its convex hull, then Test, where the slider drives the joint in the simulator, and Back |
|
In the lidar materials clip the viewer’s lidar views draw two tones, strong and attenuated, not a colour scale. The presets’ reflectivities are illustrative, not measured.
The wind tunnel clips show where the air goes on a coarse grid. Their Cd, 1.35 at 2M cells, is far from the Ahmed body’s measured 0.29 to 0.30 (docs/design/domains/aerodynamics.md §10). The stripes in its pressure panel are a known defect of the flow, not of the video (cfd.md).