README.md

Team ViGIR's Behavior Synthesis stack

ROS Python packages for automatically synthesizing high-level robot behaviors (in the form of executable SMACH-based state machines) from the user's task specifications via reactive Linear Temporal Logic (LTL) synthesis.

About

Experimental Demonstration on ATLAS

Click on the image below to watch a demonstration of Behavior Synthesis on ATLAS.

Synthesis Workflow

The nominal workflow is depicted in the diagram below:

Dependencies

For vigir_ltl_specification, you need:

• ReSpeC for compiling LTL specifications from the high-level user requests

For vigir_ltl_synthesizer, you need:

• slugs synthesizer for LTL synthesis (an installation script is provided)

For using the full stack, you also need:

• flexbe_behavior_engine for sending synthesis requests via the FlexBE app's GUI and for access to the StateInstantiation.msg
• vigir_behaviors for access to the library of states and for saving the synthesized behaviors

Publications

• Spyros Maniatopoulos, Philipp Schillinger, Vitchyr Pong, David C. Conner, and Hadas Kress-Gazit, "Reactive High-level Behavior Synthesis for an Atlas Humanoid Robot", IEEE International Conference on Robotics and Automation (ICRA), Stockholm, Sweden, May 2016 (to appear).

Maintainers:

• Spyros Maniatopoulos (@spmaniato, [email protected])

License

BSD-3 (see LICENSE file)