Repository navigation
sim2: Go2 with its L1 lidar - #4913
Draft
jeff-hykin wants to merge 92 commits into
Draft
jeff-hykin wants to merge 92 commits into
jeff-hykin wants to merge 92 commits into
Conversation
Contributor
There was a problem hiding this comment.
go2 replay realtime (arm64)
Details
| Benchmark suite | Current: b05daaf | Previous: 83800d9 | Ratio |
|---|---|---|---|
peak memory |
1762.27 MB |
1745.215 MB |
1.01 |
peak threads |
369 threads |
368 threads |
1.00 |
network (transport) |
2725.795 MB |
2730.954 MB |
1.00 |
disk write |
1.824 MB |
1.863 MB |
0.98 |
instructions |
137.106 G |
138.7 G |
0.99 |
This comment was automatically generated by workflow using github-action-benchmark.
❌ 8 Tests Failed:
View the top 3 failed test(s) by shortest run time
To view more test analytics, go to the Test Analytics Dashboard |
jeff-hykin
force-pushed
the
jeff/feat/sim2_go2_l1
branch
from
October 30, 2026 16:39
ee57f5f to
bd79aff
Compare
jeff-hykin
force-pushed
the
jeff/feat/sim2_go2_l1
branch
from
October 30, 2026 21:04
bd79aff to
045dc26
Compare
…nd actions, SimL1 reproduces the head L1's rolling voxel map, unitree-go2-sim2-l1 blueprint
jeff-hykin
force-pushed
the
jeff/feat/sim2_go2_l1
branch
from
October 31, 2026 03:02
045dc26 to
50ea95a
Compare
…s made from its MJCF (GO2_SIM_MODEL), with the L1 and camera frames the blueprint names; SimL1 takes only its frame_id; Go2Sim publishes only odom -> base_link
… config validation takes no lazy path object
…o2.xml with the rig's frames as sites (Go2Sim.l1_mount_frame, lidar_mount_frame, camera_mount_frame); unitree-go2-sim2-l1 poses mjcf_to_urdf(Go2Sim.mjcf) and its sensors name Go2Sim's frames
…vation/WorldObservations
# Conflicts: # dimos/sim2/mujoco_backend.py
…it with @shared_config(UrdfManager); Go2Sim sends its URDF
…as steady stamps, so a loaded machine's slow first point cloud does not fail it
…d's clock SimL1 defaulted to geom group 0 alone. That is the default the Livox's own config records as the one that made the beam pass through every desk, chair and robot -- and a converted scene puts collision geometry in group 3: office_555 has no group-0 geoms at all. So the L1 published an empty map every 0.13 s and nothing said why. It takes (0, 3), as the Livox does. handle_cmd_vel set its deadline from time.monotonic and _act compared against it, inside a world where every other rate paces on world.stamp(). At speed=4 a command expired after a quarter of CMD_VEL_TIMEOUT of sim time. Both ends use the world's clock now, and `time` is no longer imported here.
Measured, with three-g1-office-555 running on this machine: the live world logged `the world stopped: KeyError(... is neither a file this description names nor an asset the last scene published)` twice while this file's tests ran, each time right after a second `starting the world` in its own log. Two names collide. `SCENE_MANAGER` here was `scene/usdmanager`, which is what that blueprint's `SCENE_NAMESPACE = "scene"` makes of its own scene manager, and an unnamespaced `sim_module()` is `mujocosim` on both sides. The per-run `zenoh_domain` this already took does not help: a domain prefixes key expressions, so it keeps the *streams* apart, while an RPC goes out on the instance name. So the office's sensors asked this test's manager for a scene -- whose assets are published under different names -- and this test's teardown stopped the office's world. Only those two needed it. Every sensor in that blueprint is inside a robot's `.namespace(name)`, so it is `main/livoxsim` and never `livoxsim`; renaming the sensors here as well was tried and reverted, since it changed nothing. No behaviour change: `test_sim_time` has the same three failures with and without this, on a quiet box and on a busy one. They are `color_image: 0`, which reproduces with this file untouched on this branch and is a separate matter.
This branch was successfully deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Go2Sim walks main's Go2 policy on sim2 PD servos; SimL1 reproduces the head L1's rolling 5 cm voxel map measured from real recordings; unitree-go2-sim2-l1.