Gazebo roadmap
From Player
This is the Changelog and Roadmap of Gazebo, current development effort is heading towards getting 0.8 version out. See [Development_goals | development goals] page for long term goals.
Contents |
[edit]
Finished
[edit]
Since version 0.8-pre2
[edit]
Version 0.8-pre2 (2008.03)
- New organization of the Media
- Solved issues with 64-bit machines
- Planes can be oriented in any position
- The grid can be disabled
- Allow gazebo to run without any GUI
- Steering wheels controller
- New SVN structure
- Friction can be configured in the XML
- Shadow types and buffer index support in XML
[edit]
Version 0.8-pre1 (2007.11)
- Adoption of scons over autotools
- Modular physics engine (ODE integrated)
- OGRE rendering engine
- XML physical description of models
- Runtime pairing of interfaces and controllers
[edit]
TODO (previous to 0.8 release)
- Scons
- The current implementation has no checks for required libraries or appropriate error messages.
- Modular physics
- Need to integrate physsim into gazebo, and properly define a general purpose API for communicating with different physics engines. Surely that API will be PAL.
- OGRE HUD and better user interface
- Create a heads up display to display relevant information about the current simulation. Also allows the user to select objects in the environment, see information about them, and manipulate them. Possible even allow them to write and edit python scripts.
- Python script interface
- Someone with good expertise with Python and C is required for this. See the above section for the desired effect. Before even starting on this, it would be a good idea to have a discussion on the mailing list.
- Better support for player proxies
- Gazebo is really lacking in this department. We need to implement numerous relevant proxies.
- Distributed Gazebo
- This is a bit more out-there. But it would be really nice to run gazebo across numerous. Or run a central server that accepts client connections. Thoughts?
- Read in and use Unreal/Doom/Quake maps.
- Ogre might support this, but I haven't checked.
- Possibly interface with google maps??
- Automatically create an environment based on layout of city streets, and randomly populate with buildings.
- Multi-level buildings, elevators, doors, staircases, buttons and switches.
- 3D-audio (openal).
- WIFI simulation
- GPS simulation
- More sensors: stereo-vision, light sensors, touch, maybe even heat sensors.
- Scripting interface for all models
- Allow each model to call a user-defined python script. This needs to be flushed out much more. But the general idea is to allow formerly static objects (such as doors, lights, boxes, cameras) to be movable and controllable. This idea can also be applied to robots for special needs.

