swarmsim.sensors.Targeter#

Targeter Sensor class.

class swarmsim.sensors.Targeter.Targeter(agent=None, parent=None, distance=inf, false_negative=0.0, time_step_between_sensing=1, target_mode=None, hold='retargetonloss', attribute=None, store_history=False, show=False, seed=None, target_team=None, target_name=None, **kwargs)[source]#

Discovers agents or objects near the sensor. Can lookup target by team or name.

Parameters:

  • distance (float, default=float('inf')) – The maximum distance from the sensor to look for targets.

  • false_negative (float, default=0.0) – The probability of a target being missed.

  • time_step_between_sensing (int, default=1) – The number of time steps between sensory perceptions.

  • target_mode (str, default=None) – How to select the single target from the matching set. Can be ‘closest’, ‘furthest’, or None.

  • hold (str, default='retargetonloss') –

    When to change targets. Can be 'alwaysretarget', 'retargetonloss', 'holdlost', 'holdfirst'.

    • 'alwaysretarget' Always search and acquire closest/farthest target,

      regardless of whether the previous target is still in range.

    • 'retargetonloss': The sensor will return the same target until that target is no longer in range.

    • 'holdlost': Same as 'retargetonloss', but will return the previous target until a new target is found.

    • 'holdfirst': The sensor will lock onto the first target it finds and always return that target.

  • attribute (str, default=None) – If specified, the sensor will use attribute lookup on the target. If None, the target object itself is returned.

  • store_history (bool, default=False) – Whether to store the history of targets.

  • show (bool, default=False) – Whether to draw the sensory vector.

  • seed (int, default=None) – The seed for the random number generator. If None, the parent’s RNG is used.

  • target_team (str | int | None, default=None) – The team to look for targets in. If None, any team is considered.

  • target_name (str, default=None) – The name of the target to look for.

Attributes:
position

Methods

as_config_dict

determineState

draw

from_dict

fuzz

get_state

look_for_agents

select_target

set_agent

set_parent

set_seed

state_pretty

step
DEBUG = False[source]#
config_vars = ['static_position', 'n_possible_states', 'show', 'distance', 'bias', 'fn', 'time_step_between_sensing', 'target_mode', 'hold', 'attribute', 'store_history', 'show', 'target_team', 'target_name', 'seed'][source]#
determineState(agent)[source]#
draw(screen, offset=((0, 0), 1.0))[source]#
static from_dict(d)[source]#
fuzz(agent)[source]#
look_for_agents(world: RectangularWorld)[source]#
select_target(bag)[source]#
step(world, only_check_goals=False)[source]#

Classes

Targeter([agent, parent, distance, ...])

Discovers agents or objects near the sensor.