I wanted to create and then solve a PDDL plan and then carry it out with the Stretch Robot. I was wondering if there is any preexisting functionality for basic pick-and-place tasks?
Hey @JonathanDistler25, we don’t have any tutorials that demonstrate carrying out a PDDL plan on Stretch. I’d suggest piecing together a few tutorials to achieve your goal. I’ve found this PDDL tutorial pretty good: Getting Started with PDDL | LearnPDDL
On the Stretch side, I’d recommend writing a ROS2 node who’s job it is to parse the solution and call the services/topics you need. To get a sense of what primitive behaviors exist, I’d recommend checking out our Nav2 tutorial: Demo 1 - Mapping & Navigation - Stretch Docs or our Stretch AI tutorial: GitHub - hello-robot/stretch_ai
Let me know if you have any questions.
The hardware requirements for the Stretch AI tutorial suggested RE3, but does it still work for the RE1 model?
Some of the Stretch AI “apps” require the eye-in-hand camera, which isn’t present on the older RE1 model. But navigation and perception apps based on the pan/tilt head should work just fine.