def __init__(self, world_params):
# Simulation global parameters
self.world_params = world_params
# List of robots active on map
self.robots = {}
# Moves aggregated each turn
self.move_buffer = {}
# Log buffer storing text information of events
self.log = []
wm = world_params['type']
# Make simulation deterministic by setting seed
if not world_params['random']:
np.random.seed(1)
if wm == 'maze':
length = world_params['length']
width = world_params['width']
self.world_model = Maze(length, width)
else:
raise KeyError('Entered world model does not match existing')
class Simulation:
# Initialize world model based on available ones
def __init__(self, world_params):
# Simulation global parameters
self.world_params = world_params
# List of robots active on map
self.robots = {}
# Moves aggregated each turn
self.move_buffer = {}
# Log buffer storing text information of events
self.log = []
wm = world_params['type']
# Make simulation deterministic by setting seed
if not world_params['random']:
np.random.seed(1)
if wm == 'maze':
length = world_params['length']
width = world_params['width']
self.world_model = Maze(length, width)
else:
raise KeyError('Entered world model does not match existing')
# Add robot to model at some random empty position
def add_robot(self, robot_params, init_state):
rname = robot_params['type']
robot_id = robot_params['id']
if rname == 'car':
if not init_state:
init_state = np.append(self.get_empty_pos(),0)
robot = Car(robot_params, init_state)
else:
raise KeyError('Entered robot model does not match existing')
self.robots[robot_id] = robot
def add_sensor(self, robot_id, sensor_params):
self.robots[robot_id].add_sensor(sensor_params)
# Return empty state on map where a robot can be placed
# Need to check that other robot states do no conflict
def get_empty_pos(self):
occupied = True
while occupied:
wm_empty = self.world_model.get_empty_pos()
# Check if another robot occupies this state
occupied = False
for robot_id in self.robots:
if self.robots[robot_id].get_position() == wm_empty:
print 'Issue: get_empty_pos - state occupied'
occupied = True
return wm_empty
# Return dictionary of state variables
def get_state(self):
# Map of the world
world = self.world_model.get_map()
# States/Poses of all robots on field
robots = self.robots
#Sensor readings for all robots on field
sensors = self.read_sensors()
data = {'world':world,'robots':robots,'sensors':sensors}
return data
# Returns dictionary of sensor readings with robot id being the key
def read_sensors(self):
readings = {}
occupied = self.get_occupied()
for robot_id in self.robots:
read = self.robots[robot_id].get_sensor_readings(occupied)
readings[robot_id] = read
return readings
# Return a matrix with ones where the space is occupied by either
# an obstacle or another robot
# Currently only returns matrix with world model obstacles
def get_occupied(self):
return self.world_model.get_map()
# add move to the buffer to be executed at the end of the turn
def add_action(self, action, robot_id):
if robot_id in self.move_buffer:
raise Exception('Each robot can only take one move per turn')
self.move_buffer[robot_id] = action
# Update state based on moves - moves taken based on FIFO order
def update_turn(self):
occupied= self.get_occupied();
for robot_id in self.move_buffer:
action = self.move_buffer[robot_id]
# need to make sure wm is passed as reference
self.robots[robot_id].make_move(action, occupied)
self.move_buffer = {}
