Overview

The behavior of the robot is mainly controlled in the node node. Here it is decided which actions we would like to take. This is organized in two steps:

Action Collection

In this step, the world state and other inputs are used to create a priority sorted list of actions, which we would like to take.

Example

let mut actions = vec![
            Action::Unstiff,
            Action::SitDown,
            Action::Penalize,
            Action::Initial,
            Action::FallSafely,
            Action::StandUp,
        ];

Depending on the current situation, different actions are added to the list of available actions.

Example

If the current role is keeper and a penalty shootout is happening, the actions Jump and PrepareJump are added.

    match world_state.robot.role {
        Role::Keeper => match world_state.filtered_game_controller_state {
            Some(FilteredGameControllerState {
                game_phase: GamePhase::PenaltyShootout { .. },
                ..
            }) => {
                actions.push(Action::Jump);
                actions.push(Action::PrepareJump);
            }
            _ => actions.push(Action::DefendGoal),
        },
        ...

Action Selection

Now, the list of actions is iterated until an action is found, which is executable. This action returns a so-called motion_command, which is handed over to the motion_selector in motion.

LED Eyes Documentation

Left Eye

Multiple different things in the following order:

1. Red: in top or bottom half of the eye, latest processed image is longer than 1 second ago of respectively top or bottom camera (vision cycler is stalled/crashed/restarting)
2. Yellow: Referee Ready or FreeKick pose detected
3. Purple: Referee Ready or FreeKick pose percepted this cycle
4. Green: Ball Percept this control cycle
5. Black

Right Eye

Always based on current role:

• Blue: Defender, Midfielder
• Yellow: Keeper, ReplacementKeeper
• Black (off): Loser
• White: Searcher
• Red: Striker
• Turquoise: StrikerSupporter