Contents Menu Expand Light mode Dark mode Auto light/dark, in light mode Auto light/dark, in dark mode Skip to content
Lagrange
Logo

Overview

  • Robots and worlds
  • Feature videos

User guide

  • User guide
  • Installing Lagrange
  • Tutorial: an AMR in a warehouse
  • Tutorial: a quadcopter, from hover to a PX4 mission
  • Tutorial: test your own drone, step by step
  • Tutorial: a robot’s power system from a schematic
  • Tutorial: weather, interference and radio links
  • Tutorial: huge worlds
  • Tutorial: the sample arm with MoveIt 2
  • Tutorial: a car in a city block
  • Tutorial: people around your robot
  • Tutorial: heat in the parts, and a robot’s own interference
  • Autoware with a simulated car
  • The run file
  • The robot builder
  • The viewer’s workspaces
  • Concepts
  • Profiles
  • Sensors
  • The lagrange command
  • Writing a module
  • Writing a plugin

Physics reference

  • Physics reference
  • Rigid body
  • Vehicles and wheeled bases
  • Water
  • Rotors
  • Downwash
  • Aerodynamics: drag, wind and the wind tunnel
  • Motors
  • Battery
  • The electrical network: buses, converters and the parts between them
  • Thermal
  • Sensors
  • Weather
  • EMI and radio links
  • Traffic: cars, lights, crossings and scenarios
  • Humans: people who walk
  • Tethers
  • Heat conduction in solids

Validation

  • Validation report (v1.0, brought up to date 2026-11-03)
  • Lagrange compared with Gazebo, MuJoCo, PyBullet and others

Design

  • Domain design documents
  • Domain: aerodynamics, wind and rotor aero
  • CFD: methods, equations and what was measured
  • Domain: comms
  • Domain: electrical
  • Domain: electromagnetic interference (EMI)
  • Domain: hydrodynamics (marine)
  • Domain: radiation
  • Domain: rigid body and articulations
  • Domain: rotor downwash and outwash
  • Domain: sensors
  • Domain: thermal
  • Domain: traffic
  • Domain: vehicle dynamics
  • Domain: weather
  • Architecture
  • Formats
  • UI↔server protocol
  • Viewer
  • Huge worlds
  • Robot classes
  • Security review (v1.0)
  • Performance benchmarks
  • Dependencies

Releases

  • Lagrange v1.4.0: build a robot in the viewer (release notes)
  • Lagrange v1.3.0: 37 new sensors, a faster simulator, hardened protocol (release notes)
  • Lagrange v1.2.1: one command to install (release notes)
  • Lagrange v1.2.0: cables, heat in the parts, a robot’s own interference (release notes)
  • Lagrange v1.1.0: easier to use, checked against measurements (release notes)
  • Lagrange v1.0.0: stable (release notes)
  • Lagrange v0.7.0: huge worlds (release notes)
  • Lagrange v0.6.0: cars and autonomous cars (release notes)
  • Lagrange v0.5.0: weather, interference and radio links (release notes)
  • Lagrange v0.4.0: drones (release notes)
  • Lagrange v0.3.0: arms, electrical and thermal (release notes)
  • v0.2.0 compared with Gazebo Jetty: the sample AMR
  • Lagrange v0.2.0: simulation core and AMR (release notes, draft)
Back to top
View this page
Edit this page

Lagrange v0.5.0: weather, interference and radio links (release notes)¶

v0.4 added drones. v0.5 adds the world around the robot:

  • the site’s atmosphere, with fog, rain and dust that blind lidars and cameras and wet the floor;

  • volumetric fields of fog and wind, including wind imported from OpenFOAM;

  • electromagnetic interference from power lines, drives, welders and transmitters;

  • radio links and CAN buses whose losses reach the robot’s ROS 2 topics.

Status: released on 2026-10-25 as v0.5.0, at the owner’s request.

Highlights¶

  • Weather (lagrange_weather, domains/weather.md §8).

    • The site has no defaults. Temperature, pressure and humidity come from the profile, --weather or a WeatherAPI_1 prim.

    • Moist air density.

    • Fog: Koschmieder at 550 nm, Kim’s wavelength exponent for 905 and 1550 nm.

    • Rain: Marshall–Palmer. Dust: quartz spheres.

    • Behavioural wet friction (applied by both physics backends) and lens contamination.

    • The clear substitute.

  • Sensors in weather.

    • The lidar’s range follows the lidar equation through the extinction and a dirty window.

    • False returns from fog backscatter and from raindrops, on their own random stream.

    • The camera’s Koschmieder veil.

    • All per beam on the CPU, on every tier.

  • Fields (lagrange_fields, NanoVDB 10.0.1, MPL-2.0).

    • Fog and wind grids, compressed to a few kB.

    • The lidar integrates fog along each beam, the camera per block of pixels, and link drag and rotors fly in the wind grid.

    • The viewer’s Environment workspace gets the real weather and slices of fog, wind and |B|.

    • Samples: a fog bank in the warehouse and a sheared wind with a wake over the open field.

  • Wind from CFD. lagrange_foam_to_vdb turns an OpenFOAM case’s ASCII cell centres and velocities into a wind grid. OpenFOAM is not a dependency.

  • EMI (lagrange_emi, domains/emi.md §8).

    • Sources:

      • power lines, as straight conductors with harmonics;

      • motor drives, as dipoles driven by their motor’s current;

      • welders;

      • transmitters, by their far field;

      • field maps.

    • Victims: magnetometers add the field at their own rate, so a 60 Hz line aliases, and IMUs add noise.

    • Encoder glitches (applied to ros2_control’s encoders), CAN bit-error rate, radio interference power and harness voltage.

    • Grid frequencies are site parameters. The zero substitute.

  • Comms (lagrange_comms, domains/comms.md §8).

    • Radio links: log-distance path loss, walls of the fixed scenery, correlated shadowing, noise plus EMI interference, AWGN bit errors for BPSK to 64-QAM, and retries.

    • CAN fault confinement by ISO 11898-1.

    • --link-topics makes the ROS 2 bridge lose and delay mapped topics, /cmd_vel included.

    • The ideal substitute.

  • Schemas.

    • New: WeatherAPI_1, EmiSourceAPI_1, EmiSusceptibilityAPI_1, SensorMaterialAPI_1, RadioLinkAPI_1, CanBusAPI_1.

    • The loader places every prim on its body, so sources, victims and radios move with the robot.

  • Command line. --weather, --link-topics, and --domain weather|emi|comms=off. Units s and bit/s.

  • Samples.

    • Profiles: amr_fog.yaml, quad_x500_power_line.yaml.

    • Robots: quad_x500_emi.usda, amr_diffdrive_wifi.usda with its link map.

    • World: power_line.usda.

    • Fields: assets/fields/*.nvdb.

  • Docs. The environment tutorial (docs/guide/tutorial-environment.md, run by tools/ci/tutorial_environment.sh).

Acceptance (roadmap §6)¶

Epic

Acceptance

Status

Evidence

Weather module

domains/weather.md §7

partly met

Koschmieder and Beer–Lambert exact, dust, moist air (test_weather). Rain rate recovered within 20 %, not 5 %: Marshall–Palmer with Atlas’s fall speeds is that loose, and the test states it. Wet friction applied by MuJoCo and Jolt; the literature range check is owed

Sensor weather effects

Detection-range validation

partly met

The lidar equation, clutter trends and Koschmieder’s veil are tested (test_weather_effects), and end to end in the server. Calibration against fog- and rain-chamber data (DENSE) is owed; the fog clutter gain is uncalibrated. GPU kernels were not needed (per-beam CPU on every tier)

Fields infrastructure

Analytic-field tests; NanoVDB version (R10)

met, with owed items

test_fields (trilinear exactness, optical depth, vector shear); NanoVDB 10.0.1 chosen. GPU sampling is not needed yet. The viewer shows 2D slices, not 3D volumes

EMI module

domains/emi.md §7

partly met

Conductor, dipole and far-field closed forms at 1e-6; the 60 → 40 Hz alias exact (test_emi); the quadcopter’s magnetometer reads 8.08 µT RMS at 9.9 m from 400 A, as μ0·I/(2πd) gives. Owed: published field measurements near lines; the end-to-end recording (Q16)

Comms module

domains/comms.md §7

partly met

Friis, k·T·B, the BPSK 1e-5 point, exact CAN transitions, wall counting and the shadowing spread (test_comms); topics lost and delayed in the ROS 2 bridge. Owed: ITU-R P.1238 parameters, the RSSI recording (Q16), the MAVLink bridge

Offline CFD import

Round trip of an analytic field

met

test_fields’ Foam tests: a linear wind through OpenFOAM-format files within 1e-5; coarser voxels average to the voxel centre. A real case’s files are not checked yet (no OpenFOAM install)

Schemas

RFC 0002 amendments accepted

waiting for the owner

Merged under the v0.5 mandate (progress.md); the owner’s review is open

Known issues¶

  • Not applied yet: CAN errors on any bus traffic but the modelled node’s, and the MAVLink bridge’s telemetry over a link. SensorMaterialAPI_1 is defined but no sensor reads it.

  • Uncalibrated: the lidar’s fog clutter gain, and the indoor radio parameters of the samples.

  • NanoVDB is header-only (libnanovdb-dev). Until it is installed on a machine, the build uses a local copy of the headers (NANOVDB_INCLUDE_DIR or ~/.local/nanovdb).

  • Start-up warnings. The Python API and lagrange validate build a run before they apply a profile, so they still log the missing site parameters that the profile then gives.

  • From earlier releases: Jolt’s drive-wheel slip, Nav2’s occasional start-up discovery failure, rotor strobing in the viewer.

  • Not run while GitHub Actions are disabled: the sanitizer builds and the nightly jobs. Everything else was verified locally with colcon: 811 tests above lagrange_core, 0 failures.

Release checklist¶

  • [x] Every v0.5 epic done or partly done, with evidence above

  • [x] Release notes drafted

  • [x] Gated changes merged under the v0.5 mandate listed for review (progress.md)

  • [ ] Owner review of those merges (the weather, EMI, sensor-material and comms schemas)

  • [ ] GitHub Actions re-enabled and green on dev

  • [x] Released: dev merged into main and tagged v0.5.0 on 2026-10-25 at the owner’s request (release-please could not run with GitHub Actions disabled)

Next
Lagrange v0.4.0: drones (release notes)
Previous
Lagrange v0.6.0: cars and autonomous cars (release notes)
Copyright © Open Physical Systems contributors, Apache-2.0
Made with Sphinx and @pradyunsg's Furo
On this page
  • Lagrange v0.5.0: weather, interference and radio links (release notes)
    • Highlights
    • Acceptance (roadmap §6)
    • Known issues
    • Release checklist