-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtraffic_light_control.py
392 lines (343 loc) · 17.1 KB
/
traffic_light_control.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
'''
Author : Mojtaba Sajjadi
'''
from __future__ import absolute_import
from __future__ import print_function
from sumolib import checkBinary
import os
import sys
import optparse
import subprocess
import random
import traci
import random
import numpy as np
import keras
import h5py
from collections import deque
from keras.layers import Input, Conv2D, Flatten, Dense
from keras.models import Model
class DQNAgent:
def __init__(self):
self.gamma = 0.95 # discount rate
self.epsilon = 0.1 # exploration rate
self.learning_rate = 0.0002
self.memory = deque(maxlen=200)
self.model = self._build_model()
self.action_size = 2
def _build_model(self):
# Neural Net for Deep-Q learning Model
input_1 = Input(shape=(12, 12, 1))
x1 = Conv2D(16, (4, 4), strides=(2, 2), activation='relu')(input_1)
x1 = Conv2D(32, (2, 2), strides=(1, 1), activation='relu')(x1)
x1 = Flatten()(x1)
input_2 = Input(shape=(12, 12, 1))
x2 = Conv2D(16, (4, 4), strides=(2, 2), activation='relu')(input_2)
x2 = Conv2D(32, (2, 2), strides=(1, 1), activation='relu')(x2)
x2 = Flatten()(x2)
input_3 = Input(shape=(2, 1))
x3 = Flatten()(input_3)
x = keras.layers.concatenate([x1, x2, x3])
x = Dense(128, activation='relu')(x)
x = Dense(64, activation='relu')(x)
x = Dense(2, activation='linear')(x)
model = Model(inputs=[input_1, input_2, input_3], outputs=[x])
model.compile(optimizer=keras.optimizers.RMSprop(
lr=self.learning_rate), loss='mse')
return model
def remember(self, state, action, reward, next_state, done):
self.memory.append((state, action, reward, next_state, done))
def act(self, state):
if np.random.rand() <= self.epsilon:
return random.randrange(self.action_size)
act_values = self.model.predict(state)
return np.argmax(act_values[0]) # returns action
def replay(self, batch_size):
minibatch = random.sample(self.memory, batch_size)
for state, action, reward, next_state, done in minibatch:
target = reward
if not done:
target = (reward + self.gamma *
np.amax(self.model.predict(next_state)[0]))
target_f = self.model.predict(state)
target_f[0][action] = target
self.model.fit(state, target_f, epochs=1, verbose=0)
def load(self, name):
self.model.load_weights(name)
def save(self, name):
self.model.save_weights(name)
class SumoIntersection:
def __init__(self):
# we need to import python modules from the $SUMO_HOME/tools directory
try:
sys.path.append(os.path.join(os.path.dirname(
__file__), '..', '..', '..', '..', "tools")) # tutorial in tests
sys.path.append(os.path.join(os.environ.get("SUMO_HOME", os.path.join(
os.path.dirname(__file__), "..", "..", "..")), "tools")) # tutorial in docs
from sumolib import checkBinary # noqa
except ImportError:
sys.exit(
"please declare environment variable 'SUMO_HOME' as the root directory of your sumo installation (it should contain folders 'bin', 'tools' and 'docs')")
def generate_routefile(self):
random.seed(42) # make tests reproducible
N = 3600 # number of time steps
# demand per second from different directions
pH = 1. / 7
pV = 1. / 11
pAR = 1. / 30
pAL = 1. / 25
with open("input_routes.rou.xml", "w") as routes:
print('''<routes xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="http://sumo.dlr.de/xsd/routes_file.xsd">
<vType id="SUMO_DEFAULT_TYPE" accel="0.8" decel="4.5" sigma="0" length="5" minGap="2" maxSpeed="70"/>
<route id="always_right" edges="1fi 1si 4o 4fi 4si 2o 2fi 2si 3o 3fi 3si 1o 1fi"/>
<route id="always_left" edges="3fi 3si 2o 2fi 2si 4o 4fi 4si 1o 1fi 1si 3o 3fi"/>
<route id="horizontal" edges="2fi 2si 1o 1fi 1si 2o 2fi"/>
<route id="vertical" edges="3fi 3si 4o 4fi 4si 3o 3fi"/>
''', file=routes)
lastVeh = 0
vehNr = 0
for i in range(N):
if random.uniform(0, 1) < pH:
print(' <vehicle id="right_%i" type="SUMO_DEFAULT_TYPE" route="horizontal" depart="%i" />' % (
vehNr, i), file=routes)
vehNr += 1
lastVeh = i
if random.uniform(0, 1) < pV:
print(' <vehicle id="left_%i" type="SUMO_DEFAULT_TYPE" route="vertical" depart="%i" />' % (
vehNr, i), file=routes)
vehNr += 1
lastVeh = i
if random.uniform(0, 1) < pAL:
print(' <vehicle id="down_%i" type="SUMO_DEFAULT_TYPE" route="always_left" depart="%i" color="1,0,0"/>' % (
vehNr, i), file=routes)
vehNr += 1
lastVeh = i
if random.uniform(0, 1) < pAR:
print(' <vehicle id="down_%i" type="SUMO_DEFAULT_TYPE" route="always_right" depart="%i" color="1,0,0"/>' % (
vehNr, i), file=routes)
vehNr += 1
lastVeh = i
print("</routes>", file=routes)
def get_options(self):
optParser = optparse.OptionParser()
optParser.add_option("--nogui", action="store_true",
default=False, help="run the commandline version of sumo")
options, args = optParser.parse_args()
return options
def getState(self):
positionMatrix = []
velocityMatrix = []
cellLength = 7
offset = 11
speedLimit = 14
junctionPosition = traci.junction.getPosition('0')[0]
vehicles_road1 = traci.edge.getLastStepVehicleIDs('1si')
vehicles_road2 = traci.edge.getLastStepVehicleIDs('2si')
vehicles_road3 = traci.edge.getLastStepVehicleIDs('3si')
vehicles_road4 = traci.edge.getLastStepVehicleIDs('4si')
for i in range(12):
positionMatrix.append([])
velocityMatrix.append([])
for j in range(12):
positionMatrix[i].append(0)
velocityMatrix[i].append(0)
for v in vehicles_road1:
ind = int(
abs((junctionPosition - traci.vehicle.getPosition(v)[0] - offset)) / cellLength)
if(ind < 12):
positionMatrix[2 - traci.vehicle.getLaneIndex(v)][11 - ind] = 1
velocityMatrix[2 - traci.vehicle.getLaneIndex(
v)][11 - ind] = traci.vehicle.getSpeed(v) / speedLimit
for v in vehicles_road2:
ind = int(
abs((junctionPosition - traci.vehicle.getPosition(v)[0] + offset)) / cellLength)
if(ind < 12):
positionMatrix[3 + traci.vehicle.getLaneIndex(v)][ind] = 1
velocityMatrix[3 + traci.vehicle.getLaneIndex(
v)][ind] = traci.vehicle.getSpeed(v) / speedLimit
junctionPosition = traci.junction.getPosition('0')[1]
for v in vehicles_road3:
ind = int(
abs((junctionPosition - traci.vehicle.getPosition(v)[1] - offset)) / cellLength)
if(ind < 12):
positionMatrix[6 + 2 -
traci.vehicle.getLaneIndex(v)][11 - ind] = 1
velocityMatrix[6 + 2 - traci.vehicle.getLaneIndex(
v)][11 - ind] = traci.vehicle.getSpeed(v) / speedLimit
for v in vehicles_road4:
ind = int(
abs((junctionPosition - traci.vehicle.getPosition(v)[1] + offset)) / cellLength)
if(ind < 12):
positionMatrix[9 + traci.vehicle.getLaneIndex(v)][ind] = 1
velocityMatrix[9 + traci.vehicle.getLaneIndex(
v)][ind] = traci.vehicle.getSpeed(v) / speedLimit
light = []
if(traci.trafficlight.getPhase('0') == 4):
light = [1, 0]
else:
light = [0, 1]
position = np.array(positionMatrix)
position = position.reshape(1, 12, 12, 1)
velocity = np.array(velocityMatrix)
velocity = velocity.reshape(1, 12, 12, 1)
lgts = np.array(light)
lgts = lgts.reshape(1, 2, 1)
return [position, velocity, lgts]
if __name__ == '__main__':
sumoInt = SumoIntersection()
# this script has been called from the command line. It will start sumo as a
# server, then connect and run
options = sumoInt.get_options()
if options.nogui:
#if True:
sumoBinary = checkBinary('sumo')
else:
sumoBinary = checkBinary('sumo-gui')
sumoInt.generate_routefile()
# Main logic
# parameters
episodes = 2000
batch_size = 32
tg = 10
ty = 6
agent = DQNAgent()
try:
agent.load('Models/reinf_traf_control.h5')
except:
print('No models found')
for e in range(episodes):
# DNN Agent
# Initialize DNN with random weights
# Initialize target network with same weights as DNN Network
#log = open('log.txt', 'a')
step = 0
waiting_time = 0
reward1 = 0
reward2 = 0
total_reward = reward1 - reward2
stepz = 0
action = 0
traci.start([sumoBinary, "-c", "cross3ltl.sumocfg", '--start'])
traci.trafficlight.setPhase("0", 0)
traci.trafficlight.setPhaseDuration("0", 200)
while traci.simulation.getMinExpectedNumber() > 0 and stepz < 7000:
traci.simulationStep()
state = sumoInt.getState()
action = agent.act(state)
light = state[2]
if(action == 0 and light[0][0][0] == 0):
# Transition Phase
for i in range(6):
stepz += 1
traci.trafficlight.setPhase('0', 1)
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
for i in range(10):
stepz += 1
traci.trafficlight.setPhase('0', 2)
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
for i in range(6):
stepz += 1
traci.trafficlight.setPhase('0', 3)
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
# Action Execution
reward1 = traci.edge.getLastStepVehicleNumber(
'1si') + traci.edge.getLastStepVehicleNumber('2si')
reward2 = traci.edge.getLastStepHaltingNumber(
'3si') + traci.edge.getLastStepHaltingNumber('4si')
for i in range(10):
stepz += 1
traci.trafficlight.setPhase('0', 4)
reward1 += traci.edge.getLastStepVehicleNumber(
'1si') + traci.edge.getLastStepVehicleNumber('2si')
reward2 += traci.edge.getLastStepHaltingNumber(
'3si') + traci.edge.getLastStepHaltingNumber('4si')
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
if(action == 0 and light[0][0][0] == 1):
# Action Execution, no state change
reward1 = traci.edge.getLastStepVehicleNumber(
'1si') + traci.edge.getLastStepVehicleNumber('2si')
reward2 = traci.edge.getLastStepHaltingNumber(
'3si') + traci.edge.getLastStepHaltingNumber('4si')
for i in range(10):
stepz += 1
traci.trafficlight.setPhase('0', 4)
reward1 += traci.edge.getLastStepVehicleNumber(
'1si') + traci.edge.getLastStepVehicleNumber('2si')
reward2 += traci.edge.getLastStepHaltingNumber(
'3si') + traci.edge.getLastStepHaltingNumber('4si')
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
if(action == 1 and light[0][0][0] == 0):
# Action Execution, no state change
reward1 = traci.edge.getLastStepVehicleNumber(
'4si') + traci.edge.getLastStepVehicleNumber('3si')
reward2 = traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('1si')
for i in range(10):
stepz += 1
reward1 += traci.edge.getLastStepVehicleNumber(
'4si') + traci.edge.getLastStepVehicleNumber('3si')
reward2 += traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('1si')
traci.trafficlight.setPhase('0', 0)
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
if(action == 1 and light[0][0][0] == 1):
for i in range(6):
stepz += 1
traci.trafficlight.setPhase('0', 5)
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
for i in range(10):
stepz += 1
traci.trafficlight.setPhase('0', 6)
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
for i in range(6):
stepz += 1
traci.trafficlight.setPhase('0', 7)
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
reward1 = traci.edge.getLastStepVehicleNumber(
'4si') + traci.edge.getLastStepVehicleNumber('3si')
reward2 = traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('1si')
for i in range(10):
stepz += 1
traci.trafficlight.setPhase('0', 0)
reward1 += traci.edge.getLastStepVehicleNumber(
'4si') + traci.edge.getLastStepVehicleNumber('3si')
reward2 += traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('1si')
waiting_time += (traci.edge.getLastStepHaltingNumber('1si') + traci.edge.getLastStepHaltingNumber(
'2si') + traci.edge.getLastStepHaltingNumber('3si') + traci.edge.getLastStepHaltingNumber('4si'))
traci.simulationStep()
new_state = sumoInt.getState()
reward = reward1 - reward2
agent.remember(state, action, reward, new_state, False)
# Randomly Draw 32 samples and train the neural network by RMS Prop algorithm
if(len(agent.memory) > batch_size):
agent.replay(batch_size)
mem = agent.memory[-1]
del agent.memory[-1]
agent.memory.append((mem[0], mem[1], reward, mem[3], True))
#log.write('episode - ' + str(e) + ', total waiting time - ' +
# str(waiting_time) + ', static waiting time - 338798 \n')
#log.close()
print('episode - ' + str(e) + ' total waiting time - ' + str(waiting_time))
#agent.save('reinf_traf_control_' + str(e) + '.h5')
traci.close(wait=False)
sys.stdout.flush()