Components & libraries
WWW-POSIM combines open-source simulation and algorithm libraries with private application code to provide a web workspace, an installed app and service endpoints.
Building WWW-POSIM
Open-source components
- POSIMSimulation engine, marine models and sensors
- POSIM-TerrainGeographic terrain, meshes and imagery
- POSIM-RoutingNavigation and terrain-checked route planning
- POSIM-ControlPropulsion and motion policies
- POSIM-SLAMSensor-based estimation and mapping libraries
Private application components
- Workspace and deliveryBrowser interface, scene views and video streaming
- Accounts and sessionsSign-in, permissions, usage limits and queues
- Worlds and saved assetsPreparation jobs, vehicle library and storage
- Voyage orchestrationItinerary state, checkpoints and recovery coordination
- Desktop integrationLauncher, runtime preparation and packaging
POSIM
POSIMPOSIM-Terrain
POSIM-TerrainPOSIM-Routing
POSIM-RoutingPOSIM-Control
POSIM-ControlPOSIM-SLAM
POSIM-SLAMWWW-POSIM
WWW-POSIM — private repository; access required
The private components belong to the WWW-POSIM application repository. They connect the libraries to user accounts, persistent workspaces and the application interfaces. Available sensors and estimation modes follow the installed runtime and its providers.
Documentation by component
| Component | Usage | Technical documentation |
|---|---|---|
| Terrain | Generate terrain for your site | Terrain generation algorithms |
| Routing and control | LIVE world voyage, Control an installed workspace | Route planning & control |
| Engine and application | Workspace basics, Custom worlds | Simulation, rendering & sensors |
The terrain library generates numeric terrain, imagery and meshes. WWW-POSIM coordinates runtime tile residency, Terrain LOD: detail near the vehicle, collision readiness and session caches. A public library's availability is separate from having its runnable sensor or SLAM provider installed on the current host.
The technical series is a catalogue for maritime robotics research: Marine dynamics & rendering, Sensor models & parameters, Terrain generation algorithms, Route planning & control, Simulation, rendering & sensors, Physics-based sonar preview and GNSS & navigation. Each separates its model, configuration, outputs and interpretation limits.
Public libraries
| Library | Use and structure |
|---|---|
| POSIM | Simulation engine, marine environments, vehicle models and sensors built on Gazebo |
| POSIM-Terrain | Python geographic sampling, land/ocean source selection, mesh/SDF export and visual textures |
| POSIM-Routing | Python global, coastal and local route planners; depends on POSIM-Terrain |
| POSIM-Control | Python motion policies and C++ Gazebo leader/follower propulsion plugins |
| POSIM-SLAM | Independent visual, acoustic and inertial estimation libraries; application features require matching sensors and providers |
POSIM-Terrain
POSIM-Terrain turns a geographic region into land and seabed geometry. It combines available elevation and bathymetry sources. The numeric field supplies the terrain mesh and route collision checks; Sentinel-2 imagery supplies the land texture.
- Input: latitude, longitude, region size and grid samples
- Output: numeric elevations, meshes, SDF, textures and source manifests
- Direct use: generate a region or load a city terrain bundle
See the terrain library README for installation, satellite texture preparation and export commands.
POSIM-Routing
POSIM-Routing plans the port itinerary and routes to each arrival area. OSM land polygons guide port connections. Coastal and local planners check represented terrain before selecting a path. Port catalogue positions and water arrival targets are separate inputs.
- Input: departure, destination, passage gates and numeric terrain
- Output: global routes, coastal corridors and local waypoints
- Direct use: export a maritime route or plan a path across generated terrain
See the routing library README for route commands and planner APIs. A global route alone does not establish water depth.
POSIM-Control
POSIM-Control applies propulsion to Gazebo vehicle models. The leader follows supplied waypoints; the follower follows measured leader poses. Python policies use observed motion to set speed limits and assess stability, progress and recovery.
- Input: checked waypoints, speed requests and measured vehicle state
- Output: propulsion commands, controller status and motion policy decisions
- Direct use: use Python policies or build the native leader/follower plugins
See the control library README for model requirements, plugin builds and control topics. Native plugins require Gazebo development packages in a configured POSIM environment.
Each library has English/Korean instructions, source structure, installation commands and tests. WWW-POSIM pins the engine and libraries as Git submodules. Its runtime adapters import their implementations; packaging includes the pinned source packages.
git submodule update --init --recursive
The website and workspace services coordinate the libraries. Session ownership, accounts, video delivery, voyage checkpoints and AIS display belong to WWW-POSIM.
How they work together
POSIM-Terrain generates terrain data. POSIM-Routing uses it to plan and check paths. POSIM-Control follows approved waypoints while Gazebo computes physical motion. WWW-POSIM observes the resulting state and coordinates subsequent terrain, route and control updates.
The web workspace and installed app use the same pinned implementations. For POSIM models, sensors and engine setup, see the POSIM library guide.
City terrain examples
POSIM-Terrain examples include New York, Busan and Tokyo. Each region contains both land and ocean terrain.
- Region size: 20 × 20 km
- Numeric grid: 257 × 257 vertices
- Grid spacing: 78.125 m
- Sources: GEBCO 2026, Copernicus GLO-30 and available regional bathymetry
The bundles include meshes, SDF, numeric arrays, Sentinel-2 land textures and source manifests. Gazebo 3D world captures accompany each city. Their README explains how to regenerate them. Output grid spacing does not increase source measurement accuracy. See Terrain generation algorithms for source assumptions.
The native Mac engine includes a locally patched Gazebo Sim Sensors component under Apache-2.0. Its modified sources, build recipe, patch, notice and license are retained in www-posim/sources/gz-sim-sensors-stop inside the engine. The POSIM GNSS sensor carries its own MIT license and notice.