def startViewer(index, watchURL):
user_data_dir = args.chrome_data_dir_default + "_tmp" + str(index)
rtt_file = os.path.join(args.log_base, "rtt_%d" % index)
snapshot_file = os.path.join(args.log_base, "snapshot_%d" % index)
debug_port = 9222 + index
# Open a chrome window attached to a debugging port.
os.system('rm -r %s; cp -r %s %s' %
(user_data_dir, args.chrome_data_dir_default, user_data_dir))
cmd = ' '.join([
'%s' % args.chrome_bin,
'--remote-debugging-port=%d' % debug_port,
'--user-data-dir=%s' % user_data_dir, '--enable-devtools-experiments'
])
disp = SmartDisplay(visible=0, bgcolor='black')
disp.start()
chrome = subprocess.Popen('%s %s' % (cmd, watchURL), shell=True)
print 'Launched chrome with: %s' % cmd
options = Options()
options.add_experimental_option("debuggerAddress",
'127.0.0.1:%d' % debug_port)
driver = webdriver.Chrome(chromedriver_bin, chrome_options=options)
print '%d Launched chromedriver' % index
print '%d Loaded %s' % (index, watchURL)
time.sleep(5)
# Click the video to begin playing.
js = "document.getElementsByTagName('video')[0].click(); document.querySelector('button[data-testid=fullscreen_control]').click()"
driver.execute_script(js)
# Only one client can be attached to the remote debugging port at a time.
driver.quit()
# Remove the driver and start a viewer to record the downloaded video.
print '%d Quit chrome driver' % index
time.sleep(2)
viewer = subprocess.Popen("node record_latency.js %d %d %s %s" %
(debug_port, 20, rtt_file, snapshot_file),
shell=True)
screenshot_dir = '%s/screenshots' % args.log_base
if not os.path.exists(screenshot_dir):
os.makedirs(screenshot_dir)
count = 0
while True:
im = disp.grab()
count = count + 1
#print 'c-%d' % count
im.save('%s/%d_%d.jpg' %
(screenshot_dir, long(time.time() * 1000), count))
disp.stop()
return [chrome, viewer]
class SumoEnvironment(gym.Env):
"""
SUMO Environment for Traffic Signal Control
:param net_file: (str) SUMO .net.xml file
:param route_file: (str) SUMO .rou.xml file
:param out_csv_name: (Optional[str]) name of the .csv output with simulation results. If None no output is generated
:param use_gui: (bool) Wheter to run SUMO simulation with GUI visualisation
:param virtual_display: (Optional[Tuple[int,int]]) Resolution of a virtual display for rendering
:param begin_time: (int) The time step (in seconds) the simulation starts
:param num_seconds: (int) Number of simulated seconds on SUMO. The time in seconds the simulation must end.
:param max_depart_delay: (int) Vehicles are discarded if they could not be inserted after max_depart_delay seconds
:param delta_time: (int) Simulation seconds between actions
:param min_green: (int) Minimum green time in a phase
:param max_green: (int) Max green time in a phase
:single_agent: (bool) If true, it behaves like a regular gym.Env. Else, it behaves like a MultiagentEnv (https://github.com/ray-project/ray/blob/master/python/ray/rllib/env/multi_agent_env.py)
:sumo_seed: (int/string) Random seed for sumo. If 'random' it uses a randomly chosen seed.
:fixed_ts: (bool) If true, it will follow the phase configuration in the route_file and ignore the actions.
:sumo_warnings: (bool) If False, remove SUMO warnings in the terminal
"""
CONNECTION_LABEL = 0 # For traci multi-client support
def __init__(
self,
net_file: str,
route_file: str,
out_csv_name: Optional[str] = None,
use_gui: bool = False,
virtual_display: Optional[Tuple[int,int]] = None,
begin_time: int = 0,
num_seconds: int = 20000,
max_depart_delay: int = 100000,
time_to_teleport: int = -1,
delta_time: int = 5,
yellow_time: int = 2,
min_green: int = 5,
max_green: int = 50,
single_agent: bool = False,
sumo_seed: Union[str,int] = 'random',
fixed_ts: bool = False,
sumo_warnings: bool = True,
):
self._net = net_file
self._route = route_file
self.use_gui = use_gui
if self.use_gui:
self._sumo_binary = sumolib.checkBinary('sumo-gui')
else:
self._sumo_binary = sumolib.checkBinary('sumo')
self.virtual_display = virtual_display
assert delta_time > yellow_time, "Time between actions must be at least greater than yellow time."
self.begin_time = begin_time
self.sim_max_time = num_seconds
self.delta_time = delta_time # seconds on sumo at each step
self.max_depart_delay = max_depart_delay # Max wait time to insert a vehicle
self.time_to_teleport = time_to_teleport
self.min_green = min_green
self.max_green = max_green
self.yellow_time = yellow_time
self.single_agent = single_agent
self.sumo_seed = sumo_seed
self.fixed_ts = fixed_ts
self.sumo_warnings = sumo_warnings
self.label = str(SumoEnvironment.CONNECTION_LABEL)
SumoEnvironment.CONNECTION_LABEL += 1
self.sumo = None
if LIBSUMO:
traci.start([sumolib.checkBinary('sumo'), '-n', self._net]) # Start only to retrieve traffic light information
conn = traci
else:
traci.start([sumolib.checkBinary('sumo'), '-n', self._net], label='init_connection'+self.label)
conn = traci.getConnection('init_connection'+self.label)
self.ts_ids = list(conn.trafficlight.getIDList())
self.traffic_signals = {ts: TrafficSignal(self,
ts,
self.delta_time,
self.yellow_time,
self.min_green,
self.max_green,
self.begin_time,
conn) for ts in self.ts_ids}
conn.close()
self.vehicles = dict()
self.reward_range = (-float('inf'), float('inf'))
self.metadata = {}
self.spec = EnvSpec('SUMORL-v0')
self.run = 0
self.metrics = []
self.out_csv_name = out_csv_name
self.observations = {ts: None for ts in self.ts_ids}
self.rewards = {ts: None for ts in self.ts_ids}
def _start_simulation(self):
sumo_cmd = [self._sumo_binary,
'-n', self._net,
'-r', self._route,
'--max-depart-delay', str(self.max_depart_delay),
'--waiting-time-memory', '10000',
'--time-to-teleport', str(self.time_to_teleport)]
if self.begin_time > 0:
sumo_cmd.append('-b {}'.format(self.begin_time))
if self.sumo_seed == 'random':
sumo_cmd.append('--random')
else:
sumo_cmd.extend(['--seed', str(self.sumo_seed)])
if not self.sumo_warnings:
sumo_cmd.append('--no-warnings')
if self.use_gui:
sumo_cmd.extend(['--start', '--quit-on-end'])
if self.virtual_display is not None:
sumo_cmd.extend(['--window-size', f'{self.virtual_display[0]},{self.virtual_display[1]}'])
from pyvirtualdisplay.smartdisplay import SmartDisplay
print("Creating a virtual display.")
self.disp = SmartDisplay(size=self.virtual_display)
self.disp.start()
print("Virtual display started.")
if LIBSUMO:
traci.start(sumo_cmd)
self.sumo = traci
else:
traci.start(sumo_cmd, label=self.label)
self.sumo = traci.getConnection(self.label)
if self.use_gui:
self.sumo.gui.setSchema(traci.gui.DEFAULT_VIEW, "real world")
def reset(self, seed: Optional[int] = None, **kwargs):
if self.run != 0:
self.close()
self.save_csv(self.out_csv_name, self.run)
self.run += 1
self.metrics = []
if seed is not None:
self.sumo_seed = seed
self._start_simulation()
self.traffic_signals = {ts: TrafficSignal(self,
ts,
self.delta_time,
self.yellow_time,
self.min_green,
self.max_green,
self.begin_time,
self.sumo) for ts in self.ts_ids}
self.vehicles = dict()
if self.single_agent:
return self._compute_observations()[self.ts_ids[0]]
else:
return self._compute_observations()
@property
def sim_step(self):
"""
Return current simulation second on SUMO
"""
return self.sumo.simulation.getTime()
def step(self, action):
# No action, follow fixed TL defined in self.phases
if action is None or action == {}:
for _ in range(self.delta_time):
self._sumo_step()
else:
self._apply_actions(action)
self._run_steps()
observations = self._compute_observations()
rewards = self._compute_rewards()
dones = self._compute_dones()
info = self._compute_info()
if self.single_agent:
return observations[self.ts_ids[0]], rewards[self.ts_ids[0]], dones['__all__'], info
else:
return observations, rewards, dones, info
def _run_steps(self):
time_to_act = False
while not time_to_act:
self._sumo_step()
for ts in self.ts_ids:
self.traffic_signals[ts].update()
if self.traffic_signals[ts].time_to_act:
time_to_act = True
def _apply_actions(self, actions):
"""
Set the next green phase for the traffic signals
:param actions: If single-agent, actions is an int between 0 and self.num_green_phases (next green phase)
If multiagent, actions is a dict {ts_id : greenPhase}
"""
if self.single_agent:
if self.traffic_signals[self.ts_ids[0]].time_to_act:
self.traffic_signals[self.ts_ids[0]].set_next_phase(actions)
else:
for ts, action in actions.items():
if self.traffic_signals[ts].time_to_act:
self.traffic_signals[ts].set_next_phase(action)
def _compute_dones(self):
dones = {ts_id: False for ts_id in self.ts_ids}
dones['__all__'] = self.sim_step > self.sim_max_time
return dones
def _compute_info(self):
info = self._compute_step_info()
self.metrics.append(info)
return info
def _compute_observations(self):
self.observations.update({ts: self.traffic_signals[ts].compute_observation() for ts in self.ts_ids if self.traffic_signals[ts].time_to_act})
return {ts: self.observations[ts].copy() for ts in self.observations.keys() if self.traffic_signals[ts].time_to_act}
def _compute_rewards(self):
self.rewards.update({ts: self.traffic_signals[ts].compute_reward() for ts in self.ts_ids if self.traffic_signals[ts].time_to_act})
return {ts: self.rewards[ts] for ts in self.rewards.keys() if self.traffic_signals[ts].time_to_act}
@property
def observation_space(self):
return self.traffic_signals[self.ts_ids[0]].observation_space
@property
def action_space(self):
return self.traffic_signals[self.ts_ids[0]].action_space
def observation_spaces(self, ts_id):
return self.traffic_signals[ts_id].observation_space
def action_spaces(self, ts_id):
return self.traffic_signals[ts_id].action_space
def _sumo_step(self):
self.sumo.simulationStep()
def _compute_step_info(self):
return {
'step_time': self.sim_step,
'reward': self.traffic_signals[self.ts_ids[0]].last_reward,
'total_stopped': sum(self.traffic_signals[ts].get_total_queued() for ts in self.ts_ids),
'total_wait_time': sum(sum(self.traffic_signals[ts].get_waiting_time_per_lane()) for ts in self.ts_ids)
}
def close(self):
if self.sumo is None:
return
if not LIBSUMO:
traci.switch(self.label)
traci.close()
try:
self.disp.stop()
except AttributeError:
pass
self.sumo = None
def __del__(self):
self.close()
def render(self, mode='human'):
if self.virtual_display:
#img = self.sumo.gui.screenshot(traci.gui.DEFAULT_VIEW,
# f"temp/img{self.sim_step}.jpg",
# width=self.virtual_display[0],
# height=self.virtual_display[1])
img = self.disp.grab()
if mode == 'rgb_array':
return np.array(img)
return img
def save_csv(self, out_csv_name, run):
if out_csv_name is not None:
df = pd.DataFrame(self.metrics)
Path(Path(out_csv_name).parent).mkdir(parents=True, exist_ok=True)
df.to_csv(out_csv_name + '_conn{}_run{}'.format(self.label, run) + '.csv', index=False)
# Below functions are for discrete state space
def encode(self, state, ts_id):
phase = int(np.where(state[:self.traffic_signals[ts_id].num_green_phases] == 1)[0])
min_green = state[self.traffic_signals[ts_id].num_green_phases]
density_queue = [self._discretize_density(d) for d in state[self.traffic_signals[ts_id].num_green_phases + 1:]]
# tuples are hashable and can be used as key in python dictionary
return tuple([phase, min_green] + density_queue)
def _discretize_density(self, density):
return min(int(density*10), 9)