Simulating 2D-LiDAR in Isaac Lab

[d-corsi]

Isaac Sim Version: 4.5.0/5.0.0
Isaac Lab Version: 2.2.0

---

I tried to implement a very basic 2D LiDAR to detect obstacles around the robot—something that should be quite straightforward—but I haven’t been able to get it working correctly. After a few days of research and discussions with others, I’m starting to suspect that this might be a bug.

I implemented it as a RayCaster directly in the config file of my RL environment, like this:

@configclass
class MyEnvCfg(DirectRLEnvCfg):

    # [...] other code [...]

    #
    lidar_scanner_cfg = RayCasterCfg(
        prim_path="/World/envs/env_0/Map/GroundPlane",
        update_period=0.1,
        offset=RayCasterCfg.OffsetCfg( pos=(0.0, 0.0, 0.5), rot=(0.0, 0.0, 0.0, 0.0)),
        max_distance=10.0,
        pattern_cfg=patterns.LidarPatternCfg(
            channels=1, vertical_fov_range=[0, 0], horizontal_fov_range=[-180, 180], horizontal_res=1.0
        ),
        debug_vis=True,
        attach_yaw_only=False,
        mesh_prim_paths=["/World/envs/env_0/ObsCube"],
    )
    
    # [...] other code [...]

The LiDAR does work, and I can see the hit points rendered. However, some of the points are clearly incorrect and appear far from where they should be. Here’s a screenshot:

At first, I thought it might just be a rendering issue. But after collecting the data and plotting it manually with Matplotlib, the results matched exactly what I see on screen—confirming that the issue likely lies within the RayCaster itself:

One of the challenges is that the code is so basic and straightforward that I can’t figure out what I could have done wrong—making it really hard to debug. Do you have any suggestions?

[d-corsi]

== EDIT ==

I’ve simplified the question to highlight the core issue and updated the original post accordingly.

[jujuebaigei123]

Hello, I also encountered this problem. Have you solved it yet？

[d-corsi]

Sadly, not yet. I opened an issue here: #3013, and it looks like we’re not alone. It has been acknowledged as a bug and is currently under review.

Unfortunately, this is a significant limitation — the LiDAR is basically unusable in its current state....