def __init__(self,
tempo=1.0,
e_thresh=125,
nvect=True,
cvect=True,
svect=True,
band='bw',
concol=1.0,
snapshow=True,
ann_cycles=1,
agent_cycles=5,
act_noise=0.1,
tfile="redman4"):
EpuckBasic.__init__(self)
self.basic_setup() # defined for EpuckBasic
self.ann = AnnPuck(agent=self,
e_thresh=e_thresh,
nvect=nvect,
cvect=cvect,
svect=svect,
band=band,
snapshow=snapshow,
concol=concol,
ann_cycles=ann_cycles,
agent_cycles=agent_cycles,
act_noise=act_noise,
tfile=tfile)
self.sensors = [
self.getDistanceSensor('ps' + str(i)) for i in range(8)
]
for s in self.sensors:
s.enable(int(self.getBasicTimeStep()))
def __init__(self, tempo = 1.0, e_thresh = 125, nvect = True, cvect = True, svect = True, band = 'bw', concol = 1.0, snapshow = True,
ann_cycles = 1, agent_cycles = 5, act_noise = 0.1, tfile = "redman4"):
EpuckBasic.__init__(self)
self.basic_setup() # defined for EpuckBasic
self.ann = AnnPuck(agent = self, e_thresh = e_thresh, nvect = nvect, cvect = cvect, svect = svect, band = band, snapshow = snapshow,
concol = concol, ann_cycles = ann_cycles, agent_cycles = agent_cycles, act_noise = act_noise,
tfile = tfile, training_rounds = 50)
self.sensors = [self.getDistanceSensor('ps' + str(i)) for i in range(8)]
for s in self.sensors: s.enable(int(self.getBasicTimeStep()))
class Marvin(EpuckBasic):
def __init__(self,
tempo=1.0,
e_thresh=125,
nvect=True,
cvect=True,
svect=True,
band='bw',
concol=1.0,
snapshow=True,
ann_cycles=1,
agent_cycles=5,
act_noise=0.1,
tfile="redman4"):
EpuckBasic.__init__(self)
self.basic_setup() # defined for EpuckBasic
self.ann = AnnPuck(agent=self,
e_thresh=e_thresh,
nvect=nvect,
cvect=cvect,
svect=svect,
band=band,
snapshow=snapshow,
concol=concol,
ann_cycles=ann_cycles,
agent_cycles=agent_cycles,
act_noise=act_noise,
tfile=tfile)
self.sensors = [
self.getDistanceSensor('ps' + str(i)) for i in range(8)
]
for s in self.sensors:
s.enable(int(self.getBasicTimeStep()))
def long_run(self, steps=500):
self.ann.simsteps = steps
self.spin_angle(prims1.randab(0, 360))
self.ann.redman_run()
def enter_input(self, sensors):
#print sensors
for layer in self.ann.get_layers():
if layer.get_type() == TYPE.INPUT:
sensors = map(lambda (x): x / 4e3, sensors)
layer.init_input(sensors)
def get_output(self):
for layer in self.ann.get_layers():
if layer.get_type() == TYPE.OUTPUT:
return layer.get_output()
# User defined function for initializing and running
# the Marvin class
def run(self):
# You should insert a getDevice-like function in order to get the
# instance of a device of the robot. Something like:
# led = self.getLed('ledname')
'''
1. get_proximities(): vector with values from the 8 distance sensors
2. snapshot()
3. move, move_wheels, set_wheel_speeds
4. run_timestep, do_timed_action
'''
#self.move_wheels(left, right, duration)
#print self.get_proximities()
#pimg = self.snapshot(True)
#print "proximities: "
#print self.get_proximities()
training_data = read_training_data()
self.ann.do_training(training_data)
# Main loop
while True:
# Read the sensors:
# Enter here functions to read sensor data, like:
#for s in self.sensors:
#val = s.getValue()
sens = self.get_proximities()
self.enter_input(sens)
self.ann.execute()
output = self.get_output()
left = 0
right = 0
for i in range(len(output) / 2):
left += output[i]
right += output[len(output) - 1 - i]
self.set_wheel_speeds(left, right)
# Process sensor data here.
#self.enter_input(img)
# Enter here functions to send actuator commands, like:
# led.set(1)
# Perform a simulation step of 64 milliseconds
# and leave the loop when the simulation is over
if self.step(64) == -1: break
class Marvin (EpuckBasic):
def __init__(self, tempo = 1.0, e_thresh = 125, nvect = True, cvect = True, svect = True, band = 'bw', concol = 1.0, snapshow = True,
ann_cycles = 1, agent_cycles = 5, act_noise = 0.1, tfile = "redman4"):
EpuckBasic.__init__(self)
self.basic_setup() # defined for EpuckBasic
self.ann = AnnPuck(agent = self, e_thresh = e_thresh, nvect = nvect, cvect = cvect, svect = svect, band = band, snapshow = snapshow,
concol = concol, ann_cycles = ann_cycles, agent_cycles = agent_cycles, act_noise = act_noise,
tfile = tfile, training_rounds = 50)
self.sensors = [self.getDistanceSensor('ps' + str(i)) for i in range(8)]
for s in self.sensors: s.enable(int(self.getBasicTimeStep()))
def long_run(self,steps = 500):
self.ann.simsteps = steps
self.spin_angle(prims1.randab(0,360))
self.ann.redman_run()
def enter_input(self,sensors):
#print sensors
for layer in self.ann.get_layers():
if layer.get_type() == TYPE.INPUT:
sensors = map(lambda(x): self.scaleSensor(x), sensors)
layer.init_input(sensors)
def scaleSensor(self,x):
if(x/200 < 1):
return x/200
else:
return 0.99
def get_output(self):
for layer in self.ann.get_layers():
if layer.get_type() == TYPE.OUTPUT:
return layer.get_output()
# User defined function for initializing and running
# the Marvin class
def run(self):
# You should insert a getDevice-like function in order to get the
# instance of a device of the robot. Something like:
#print self.get_proximities()
pimg = self.snapshot(False)
print pimg
#print "proximities: "
#print self.get_proximities()
training_data = read_training_data()
self.ann.do_training(training_data)
self.ann.stop_training()
# Main loop
while True:
# Read the sensors:
sens = self.get_proximities()
# Process sensor data here.
self.enter_input(sens)
self.ann.execute()
output = self.get_output()
if self.ann.is_hardwired():
self.set_wheel_speeds(output[0]-output[2]+uniform(0.3,0.5), output[1]-output[2]+uniform(0.3,0.5))
else:
self.set_wheel_speeds(output[0], output[1])
# Enter here functions to send actuator commands, like:
# led.set(1)
# Perform a simulation step of 64 milliseconds
# and leave the loop when the simulation is over
if self.step(64) == -1: break
class Marvin (EpuckBasic):
def __init__(self, tempo = 1.0, e_thresh = 125, nvect = True, cvect = True, svect = True, band = 'bw', concol = 1.0, snapshow = True,
ann_cycles = 1, agent_cycles = 5, act_noise = 0.1, tfile = "redman4"):
EpuckBasic.__init__(self)
self.basic_setup() # defined for EpuckBasic
self.ann = AnnPuck(agent = self, e_thresh = e_thresh, nvect = nvect, cvect = cvect, svect = svect, band = band, snapshow = snapshow,
concol = concol, ann_cycles = ann_cycles, agent_cycles = agent_cycles, act_noise = act_noise,
tfile = tfile)
self.sensors = [self.getDistanceSensor('ps' + str(i)) for i in range(8)]
for s in self.sensors: s.enable(int(self.getBasicTimeStep()))
def long_run(self,steps = 500):
self.ann.simsteps = steps
self.spin_angle(prims1.randab(0,360))
self.ann.redman_run()
def enter_input(self,sensors):
#print sensors
for layer in self.ann.get_layers():
if layer.get_type() == TYPE.INPUT:
sensors = map(lambda(x): x/4e3, sensors)
layer.init_input(sensors)
def get_output(self):
for layer in self.ann.get_layers():
if layer.get_type() == TYPE.OUTPUT:
return layer.get_output()
# User defined function for initializing and running
# the Marvin class
def run(self):
# You should insert a getDevice-like function in order to get the
# instance of a device of the robot. Something like:
# led = self.getLed('ledname')
'''
1. get_proximities(): vector with values from the 8 distance sensors
2. snapshot()
3. move, move_wheels, set_wheel_speeds
4. run_timestep, do_timed_action
'''
#self.move_wheels(left, right, duration)
#print self.get_proximities()
#pimg = self.snapshot(True)
#print "proximities: "
#print self.get_proximities()
training_data = read_training_data()
self.ann.do_training(training_data)
# Main loop
while True:
# Read the sensors:
# Enter here functions to read sensor data, like:
#for s in self.sensors:
#val = s.getValue()
sens = self.get_proximities()
self.enter_input(sens)
self.ann.execute()
output = self.get_output()
left = 0
right = 0
for i in range(len(output)/2):
left += output[i]
right += output[len(output)-1-i]
self.set_wheel_speeds(left, right)
# Process sensor data here.
#self.enter_input(img)
# Enter here functions to send actuator commands, like:
# led.set(1)
# Perform a simulation step of 64 milliseconds
# and leave the loop when the simulation is over
if self.step(64) == -1: break