RAI is a multi-vendor agent framework for robotics, utilizing Langchain and ROS 2 tools to perform complex actions, defined scenarios, free interface execution, log summaries, voice interaction and more.
Is your feature request related to a problem? Please describe.
Most of main documentation is now in good form. However, parts of auxilary documentation are not actionable and serve no apparent purpose. Some other parts are outdated or inaccurate. Small parts are still missing.
Describe the solution you'd like
The smaller tasks are needed to bring documentation up in quality:
[ ] Replace "human robot interface" with "human robot interaction" for HRI.
[ ] Document internal rai_interfaces for developer docs.
[ ] Describe state retrieval and how the default works vs what developers might want to add / change.
[ ] Improve Hello RAI example in the main README - currently it misses action path and is too generic (doesn't answer the "how?" question).
[ ] Rework docs on manipulation - they serve no clear purpose currently. There should be a chapter on how to make RAI work with moveIt2, and a link to the manipulation demo.
[ ] Rework or remove docs on multimodal_messages. What is actionable for users?
[ ] Update vendors documentation. Have clear messaging about vendor support (overall and specific features).
Additional context
The task can be broken into smaller pieces or done in bulk.
Is your feature request related to a problem? Please describe. Most of main documentation is now in good form. However, parts of auxilary documentation are not actionable and serve no apparent purpose. Some other parts are outdated or inaccurate. Small parts are still missing.
Describe the solution you'd like The smaller tasks are needed to bring documentation up in quality:
rai_interfacesfor developer docs.Additional context The task can be broken into smaller pieces or done in bulk.