The bottom actions submodule
The main entrypoint is main(int argc, char** argv). Two objects are constructed when started:
The strategy selection submodule
Action selection are written in module dqn_network, using Python and Tensorflow.
directory structure
- camke_module, config, data: Config files.
- proto: Messages to communicate to other ROS nodes.
- behavior_tree: Strategy selection using the behavior tree model.
- dqn_decision: Strategy selection using the Deep Q Learning.
- blackboard: Exposes the roborts_decision::StrategyExecute class, which provides all information.
- strategy: Code to execute the chosen strategy.
- executor: Provides low-level API that move the robots.