data_gathering package

Subpackages

Submodules

data_gathering.information_manager module

Aim of this module is to provide a less convoluted access to information, i.e. distill the information from the data and return high level information

class data_gathering.information_manager.InformationManager(agent: LunaticAgent, update_information: bool = True)

Bases: object

Tracks global information, e.g. all actors, traffic lights, etc. as well as agent specific information, e.g. current speed, current location, and nearby actors in relation to the agent.

Tip

global information is attributed by ClassVar and staticmethod and is shared across all instances, this information is constant for the current tick.

Parameters:
relevant_traffic_light: carla.TrafficLight | None = None

Result of CarlaDataProvider.get_next_traffic_light()

relevant_traffic_light_distance: float = inf

Distance to the relevant_traffic_light Is float(‘inf’) if relevant_traffic_light is None.

gathered_information: InformationManager.Information

InformationManager.Information gathered during tick()

vehicles: ClassVar['list[carla.Vehicle]']

List of all tracked vehicles

walkers: ClassVar['list[carla.Walker]']

List of all tracked pedestrians

static_obstacles: ClassVar['list[carla.Actor]']

List of all tracked static obstacles

obstacles: ClassVar['list[carla.Actor]']

Union of vehicles, py:attr:walkers and py:attr:static_obstacles

lights_map: ClassVar['Dict[int, carla.Waypoint]'] = {}

Map of traffic lights to their trigger waypoints

frame: ClassVar['int | None'] = None

Last frame the InformationManager was updated.

The current frame of the world should be passed to the global_tick method.

__init__(agent: LunaticAgent, update_information: bool = True)
Parameters:
state_counter: Dict[AgentState, int]

Tracks different AgentState and the amount of ticks the agent is in this state.

detect_next_traffic_light()

Set the relevant_traffic_light and relevant_traffic_light_distance if not set.

Note

Does not check for planned path but current route along waypoints, might not be exact.

This function is automatically called in :py:meth:`tick`

tick()

Tick the information manager and update the information for the corresponding agent.

class Information(current_waypoint: carla.Waypoint, current_speed: float, current_states: Dict[AgentState, int], relevant_traffic_light: carla.TrafficLight | None, relevant_traffic_light_distance: float, vehicles: List[carla.Vehicle], walkers: List[carla.Walker], static_obstacles: List[carla.Actor], obstacles: List[carla.Actor], walkers_nearby: List[carla.Walker], vehicles_nearby: List[carla.Vehicle], static_obstacles_nearby: List[carla.Actor], obstacles_nearby: List[carla.Actor], traffic_lights_nearby: List[carla.TrafficLight], distances: Dict[carla.Actor, float])

Bases: NamedTuple

Data gathered by the InformationManager which is passed to the agent.

Parameters:
current_waypoint: carla.Waypoint

Alias for field number 0

current_speed: float

Alias for field number 1

current_states: Dict[AgentState, int]

Alias for field number 2

relevant_traffic_light: carla.TrafficLight | None

Alias for field number 3

relevant_traffic_light_distance: float

Alias for field number 4

vehicles: List[carla.Vehicle]

Alias for field number 5

walkers: List[carla.Walker]

Alias for field number 6

static_obstacles: List[carla.Actor]

Filtered obstacles by InformationManager.OBSTACLE_FILTER

count(value, /)

Return number of occurrences of value.

index(value, start=0, stop=9223372036854775807, /)

Return first index of value.

Raises ValueError if the value is not present.

obstacles: List[carla.Actor]

Union of vehicles, walkers and static_obstacles

walkers_nearby: List[carla.Walker]

Alias for field number 9

vehicles_nearby: List[carla.Vehicle]

Alias for field number 10

static_obstacles_nearby: List[carla.Actor]

Alias for field number 11

obstacles_nearby: List[carla.Actor]

Alias for field number 12

traffic_lights_nearby: List[carla.TrafficLight]

Alias for field number 13

distances: Dict[carla.Actor, float]

Distances to all actors in obstacles

OBSTACLE_FILTER: str = 'static.prop.[cistmw]*'

fnmatch for obstacles that the agent will consider in its path. https://carla.readthedocs.io/en/latest/bp_library/#static

static get_traffic_lights() → Dict[carla.TrafficLight, carla.Transform]
Return type:

Dict[carla.TrafficLight, carla.Transform]

static get_trafficlight_trigger_waypoint(traffic_light: carla.TrafficLight) → carla.Waypoint

Get the location where the traffic light is triggered.

Parameters:

traffic_light (carla.TrafficLight)

Return type:

carla.Waypoint

static global_tick(frame: int | None = None) → None

Update global information that is constant for the current tick and not agent specific.

Updates:
Parameters:

frame (Optional[int]) – The id of the current frame. If None retrieves the id from the current carla.WorldSnapshot. Multiple calls with the same frame are ignored. (default: None)

Return type:

None

static get_vehicles() → List[carla.Vehicle]
Return type:

List[carla.Vehicle]

static get_walkers() → List[carla.Walker]
Return type:

List[carla.Walker]

static cleanup() → None

Resets the global information.

Return type:

None