Launch Files
These are .launch files that can be run using roslaunch:
roslaunch rktl_autonomy <launch file>
cartpole_train.launch
This launch file starts the training a model for the cartpole environment. Wraps
the cartpole_env node. For more info,
refer to its documentation.
Launch Arguments
agent_name(default:rocket_league_agent): Name of the agent.render(default:false): If true, a window for the current simulation will be shown.rate(value:30.0): Rate at which the simulation runs, in steps per second.
Nodes
cartpole_env:cartpole_envnode in therktl_autonomypackage.
rocket_league_agent.launch
This launch file starts the agent for the rocket league environment. Wraps
the rocket_league_agent node. For
more info, refer to its documentation.
Launch Arguments
plot_log(default:false): If true, a plot will be generated.weights_dir(default:'~/catkin_ws/data/rocket_league/'): Folder containing the weights of pre-trained models.weights_name(default:'model'): Which of the models to load, from the directory defined inweights_dir.
Nodes
rocket_league_agent:rocket_league_agentnode from therktl_autonomypackage.plotter:plotternode from therktl_autonomypackage. Only run if theplot_logargument is set totrue.
rocket_league_train.launch
This launch file starts the training a model for the rocket league environment.
Wraps the rocket_league_agent node
and runs the simulator.
Loads a training environment for and starts training a model.
Launch Arguments
plot_log(default:true): If true, a plot will be generated.agent_name(default:rocket_league_agent): Name of the agent.render(default:false): If true, a window for the current simulation will be shown.sim_mode(default:'ideal'): Simulation mode for training.rate(value:10.0): Rate at which the simulation runs, in steps per second. Cannot be overridden.agent_type(value:'none'): Disables the loading of an existing agent. Cannot be overridden.
Nodes
plotter:plotternode from therktl_autonomypackage. Only run if theplot_logargument is set totrue.
Included Launch Files
rktl_perception camera.launch(Included twice, once withcamera_nameset tocam2and once withcamera_nameset tocam3)rktl_launch rocket_league_sim.launch: Loads Parameters fromconfig/rocket_league.yamlin therktl_autonomypackage. Also sets~logto$(arg agent_name)/log.
snake_eval.launch
This launch file starts the agent for the snake environment. Wraps the
snake_agent node. For more info, refer to
its documentation.
Launch Arguments
plot_log(default:false): If true, a plot will be generated.render(default:true): If true, a window for the current simulation will be shown.weights(default:'~/catkin_ws/data/snake/weights’): Folder containing the weights of pre-trained models.rate(value:10.0): Rate at which the simulation runs, in steps per second. Cannot be overridden.snake_size(value:7): Initial number of segments for the snake. Cannot be overridden.arena_size(value:10): Size of the arena the snake operates in. Cannot be overridden.
Nodes
snake_env:snake_nodenode from thesnakesimpackage.snake_agent:snake_agentnode from therktl_autonomypackage.plotter:plotternode from therktl_autonomypackage. Only run if theplot_logargument is set totrue.
snake_train.launch
This launch file starts the training a model for the snake environment. Wraps
the snake_agent node. For more info,
refer to its documentation.
Launch Arguments
plot_log(default:false): If true, a plot will be generated.agent_name(default:'snake_agent'): Name of the agent.render(default:false): If true, a window for the current simulation will be shown.rate(value:10.0): Rate at which the simulation runs, in steps per second. Cannot be overridden.snake_size(value:7): Initial number of segments for the snake. Cannot be overridden.arena_size(value:10): Size of the arena the snake operates in. Cannot be overridden.
Nodes
snake_env:snake_nodenode from thesnakesimpackage.plotter:plotternode from therktl_autonomypackage. Only run if theplot_logargument is set totrue.