Source code for pomdp_py.problems.tag.models.observation_model

import pomdp_py
from pomdp_py.problems.tag.domain.observation import *
import pomdp_py.problems.tag.constants as constants




[docs]
class TagObservationModel(pomdp_py.ObservationModel):
    """In this observation model, the robot deterministically
    observes the target location when it is in the same grid cell
    as the target. Ohterwise the robot does not observe anything."""



[docs]
    def probability(self, observation, next_state, action, **kwargs):
        if next_state.robot_position == next_state.target_position:
            if observation.target_position is None:
                return constants.EPSILON
            else:
                if observation.target_position == next_state.target_position:
                    return 1.0 - constants.EPSILON
                else:
                    return constants.EPSILON
        else:
            if observation.target_position is None:
                return 1.0 - constants.EPSILON
            else:
                return constants.EPSILON





[docs]
    def sample(self, next_state, action):
        """There is no stochaisticity in the observation model"""
        if next_state.robot_position == next_state.target_position:
            return TagObservation(next_state.target_position)
        else:
            return TagObservation(None)





[docs]
    def argmax(self, next_state, action):
        return self.sample(next_state, action)