示例#1
0
    def __init__(self):
        self._seed()
        self.viewer = None  # to be used later for rendering
        self.target_pos = None
        membrane_base.init_helper(self)

        self.object = None

        # Drawlist for rendering
        self.drawlist = []

        self.prev_state = None

        # Observation Space
        # [object posx, object posy, actuator1 pos.y, ... , actuator5 pos.y, actuator1 speed.y, ... , actuator5 speed.y]
        high = np.array([np.inf] * 16)
        self.observation_space = spaces.Box(low=-high, high=high)

        # Continuous action space; one for each linear actuator (5 total)
        # action space represents velocity
        self.action_space = spaces.Box(-1, 1, (5, ))
        self.prev_shaping = None  # for reward calculation

        self.pd_control_list = []

        self._reset()
    def __init__(self):
        self._seed()
        self.viewer = None  # to be used later for rendering
        # Object to be manipulated
        self.object = None
        # Initializing other common components in the environment
        membrane_base.init_helper(self)

        self.prev_state = None

        # Observation Space
        high = np.array([np.inf] * 15)
        self.observation_space = spaces.Box(low=-high, high=high)

        # Continuous action space; one for each linear actuator (5 total)
        # action space represents velocity
        self.action_space = spaces.Box(-1, 1, (5, ))
        self.prev_shaping = None  # for reward calculation

        self._reset()