johnfarrell / rr-godot Goto Github PK

Support multiple viewports each with their own camera.

The user camera for each viewport needs to support zooming in both perspective and orthogonal views.

Allow useres to upload list of joint positions to simulate a motion plan path.

Support custom environment importing via traditional formats such as .obj, .stl, .dae

Need to be able to play/pause simulation and physics processing to allow for scene setup.

Need to be able to read in data through Robot Raconteur wires or other transmissions.

Use Gazebo .world file to import and save world/simulation settings to quickly open a simulation.

Add in a set of text boxes that allow the user to customize the transforms of any object inside the tree.

When the active gizmos are inside an object in the environment, the user is unable to click them.

Can lead to scaling an object enough to where it covers the handle, and you are unable to scale it further without moving away from the object far enough.

Need to be able to simulate various sensors (camera, depth, LIDAR) and publish the data over RR for clients to use.

Hi There,

This is an awesome project!! 😍

Does this require your Godot fork or it should be able to run with the latest Godot.Mono build?

Need to be able to display sensor data in a 3D environment such as LIDAR sensors and depth sensors.

Need to be able to simulate a dynamic environment with simple world path animations

All objects need to be displayed in a tree-structure on the left side that updates automatically.

Use a gizmo and translate/rotate/scale environment objects like traditional 3D editing environments.

Find a way to scale the gizmos independently for each viewport. May require making 4 different gizmo instances.

Entire robot description importing via .urdf files.

Need to create a plugin ecosystem to allow users to create custom functionality.
First plugin should be .sdf file importing.

Will be related to #15

When using the 4-way viewport, the user cannot click on gizmos to control them.

If you click on the environment tree and then control the camera, the environment tree will continue to receive input, folding and unfolding sections.

Currently the joint type is shown with a label underneath the name.
This takes up a lot of unneeded space, and can be changed to a graphic shown next to the name in both the object tree and the inspector.

An API needs to be exposed to allow multiple languages (the ones supported by RR) to control the simulation.

Need to be able to import .robdef files for the sensor simulation. Allows users to re-use control code.

Need to be able to read data gathered from RR or simulation and graph it such as joint angle/time.

Be able to save/load editor layouts similar to Blender. Multiple camera angles/ etc

Need to detect collisions between any 2+ objects and display where the collision takes place.