def run(self):
camera = geometric_transformation.geometric_transformation(0, 0, 0)
# open CSV file
with open(self.camerafilepath, 'r') as cameracsv:
for line in csv.reader(cameracsv, skipinitialspace=True):
# skip row with comments or empty
if len(line) == 0: continue
if line[0][0] == '#': continue
# transform text to integers
line = map(lambda x: float(x), line)
# compute fix to errors
camera_correction = self.move()
print self.getName() + ": line:", line
print self.getName(
) + ": camera_correction:", camera_correction
# apply fix to error
camera.rotate_X_axis(line[0] + line[1] - camera_correction[0])
camera.rotate_Y_axis(line[2] + line[3] - camera_correction[1])
camera.rotate_Z_axis(line[4] + line[5] - camera_correction[2])
camera.traslate(line[6] + line[7] - camera_correction[3],
line[8] + line[9] - camera_correction[4],
line[10] + line[11] - camera_correction[5])
# show stats about the current value and distance from corrections
self.stats(line, camera_correction)
cameracsv.close()
def run(self):
camera = geometric_transformation.geometric_transformation(0,0,0)
# open CSV file
with open(self.camerafilepath, 'r') as cameracsv:
for line in csv.reader(cameracsv, skipinitialspace=True):
# skip row with comments or empty
if len(line) == 0: continue
if line[0][0] == '#': continue
# transform text to integers
line = map(lambda x: float(x), line)
# compute fix to errors
camera_correction = self.move()
print self.getName() + ": line:", line
print self.getName() + ": camera_correction:", camera_correction
# apply fix to error
camera.rotate_X_axis(line[0] + line[1] - camera_correction[0])
camera.rotate_Y_axis(line[2] + line[3] - camera_correction[1])
camera.rotate_Z_axis(line[4] + line[5] - camera_correction[2])
camera.traslate(line[6] + line[7] - camera_correction[3], line[8] + line[9] - camera_correction[4], line[10] + line[11] - camera_correction[5])
# show stats about the current value and distance from corrections
self.stats(line, camera_correction)
cameracsv.close()
def main():
point = geometric_transformation.geometric_transformation(3, 5, 7)
with open(CAMERAS_DATA_FILE, 'rb') as csvfile:
data = csv.reader(csvfile, skipinitialspace=True)
for row in data:
# skip row with comments or empty
if len(row) == 0: continue
if row[0][0] == '#': continue
# transform text to integers
row = map(lambda x: float(x), row)
point.rotate_X_axis(row[0]) # rotate of value
point.rotate_X_axis(row[1]) # rotate as the error value
print point.actual_position
point.rotate_Y_axis(row[2]) # rotate on Y as value
point.rotate_Y_axis(row[3]) # rotate as the error value
print point.actual_position
point.rotate_Z_axis(row[4]) # rotate on Y as value
point.rotate_Z_axis(row[5]) # rotate as the error value
print point.actual_position
point.traslate(row[6] + row[7], row[8] + row[9], row[10] + row[11])
print point.actual_position
print "-----------------------"
def main():
point = geometric_transformation.geometric_transformation(3,5,7)
with open(CAMERAS_DATA_FILE, 'rb') as csvfile:
data = csv.reader(csvfile, skipinitialspace=True)
for row in data:
# skip row with comments or empty
if len(row) == 0: continue
if row[0][0] == '#': continue
# transform text to integers
row = map(lambda x: float(x), row)
point.rotate_X_axis(row[0]) # rotate of value
point.rotate_X_axis(row[1]) # rotate as the error value
print point.actual_position
point.rotate_Y_axis(row[2]) # rotate on Y as value
point.rotate_Y_axis(row[3]) # rotate as the error value
print point.actual_position
point.rotate_Z_axis(row[4]) # rotate on Y as value
point.rotate_Z_axis(row[5]) # rotate as the error value
print point.actual_position
point.traslate(row[6]+row[7], row[8]+row[9], row[10]+row[11])
print point.actual_position
print "-----------------------"
def feasible_moves_selection(self,
attractiveness_weight=None,
trails_weight=None):
""" statistically select wich of the feasible move is the right one to select
attractivness is the 'a priori' desiderability of that move
trail level is the 'a posteriori' desiderability of that move
"""
# set default value if not otherwise fixed
if attractiveness_weight == None:
attractiveness_weight = self.attractiveness_weight
if trails_weight == None:
trails_weight = self.trails_weight
# memorize all evaluated position with respective fitness and trail level
evaluated_positions = list()
most_feasible = (0, 0, 0, 0)
numerator = denominator = 0
# for each move:
for move in self.feasible_moves:
# set from where the object is located
gt = geometric_transformation.geometric_transformation(
self.actual_position[0], self.actual_position[1],
self.actual_position[2])
# compute what the actual choice do to the object position
evaluated_position = gt.rototraslate_on_all_axis(
move[0], move[1], move[2], move[3], move[4], move[5])
# compute the attractivness of that move (a priory move )
a_priory_desiderability = self.compute_fitness(evaluated_position)
# print "a_priory_desiderability:", a_priory_desiderability
# compute trail level on this particular move ( a posteriori move ) using hints from others ANT's
a_posteriori_desiderability = self.trails.value(evaluated_position)
# print "a_posteriori_desiderability:", a_posteriori_desiderability
# add to evaluted position every choice with it's respective fitness
evaluated_positions.append(
(evaluated_position, a_priory_desiderability,
a_posteriori_desiderability))
# Compute now in probability how good are the choices
# First we need the denominator value
for position in evaluated_positions:
# for every evaluated position get previous calculated values
a_priory_desiderability = position[1]
a_posteriori_desiderability = position[2]
# compute denominator
denominator += (
1 - attractiveness_weight) * a_priory_desiderability + (
trails_weight * a_posteriori_desiderability)
# Then we need to compute numerator and get data in probability
count = 0
for position in evaluated_positions:
numerator = (
1 - attractiveness_weight
) * a_priory_desiderability + trails_weight * a_posteriori_desiderability
# NOTE: I've skiped tabu list exclusion to evaluate the probability of a move using it in move creation
# using a short term tabu list task
# compute the probability of that move
position_probability = numerator / denominator
# As choice are created, memorize the move number of the best one !
if most_feasible[3] < position_probability:
most_feasible = position[0][0][0], position[0][1][0], position[
0][2][0], position_probability[0]
most_feasible_id = count
count = count + 1
# return the best feasible move
return self.feasible_moves[most_feasible_id]
def feasible_moves_selection(self, attractiveness_weight = None, trails_weight = None):
""" statistically select wich of the feasible move is the right one to select
attractivness is the 'a priori' desiderability of that move
trail level is the 'a posteriori' desiderability of that move
"""
# set default value if not otherwise fixed
if attractiveness_weight == None:
attractiveness_weight = self.attractiveness_weight
if trails_weight == None:
trails_weight = self.trails_weight
# memorize all evaluated position with respective fitness and trail level
evaluated_positions = list()
most_feasible = (0,0,0,0)
numerator = denominator = 0
# for each move:
for move in self.feasible_moves:
# set from where the object is located
gt = geometric_transformation.geometric_transformation(self.actual_position[0], self.actual_position[1], self.actual_position[2])
# compute what the actual choice do to the object position
evaluated_position = gt.rototraslate_on_all_axis(move[0], move[1], move[2], move[3], move[4], move[5])
# compute the attractivness of that move (a priory move )
a_priory_desiderability = self.compute_fitness(evaluated_position)
# print "a_priory_desiderability:", a_priory_desiderability
# compute trail level on this particular move ( a posteriori move ) using hints from others ANT's
a_posteriori_desiderability = self.trails.value(evaluated_position)
# print "a_posteriori_desiderability:", a_posteriori_desiderability
# add to evaluted position every choice with it's respective fitness
evaluated_positions.append((evaluated_position, a_priory_desiderability, a_posteriori_desiderability))
# Compute now in probability how good are the choices
# First we need the denominator value
for position in evaluated_positions:
# for every evaluated position get previous calculated values
a_priory_desiderability = position[1]
a_posteriori_desiderability = position[2]
# compute denominator
denominator += (1 - attractiveness_weight) * a_priory_desiderability + (trails_weight * a_posteriori_desiderability)
# Then we need to compute numerator and get data in probability
count = 0
for position in evaluated_positions:
numerator = (1 - attractiveness_weight) * a_priory_desiderability + trails_weight * a_posteriori_desiderability
# NOTE: I've skiped tabu list exclusion to evaluate the probability of a move using it in move creation
# using a short term tabu list task
# compute the probability of that move
position_probability = numerator / denominator
# As choice are created, memorize the move number of the best one !
if most_feasible[3] < position_probability:
most_feasible = position[0][0][0], position[0][1][0], position[0][2][0], position_probability[0]
most_feasible_id = count
count = count + 1
# return the best feasible move
return self.feasible_moves[most_feasible_id]
# wait for all thread are stopped
for ant in atomic_ant_colony:
ant.join()
#
for ant in atomic_ant_colony:
ant.compute_fitness()
with open(REAL_POSITION_FILE, 'rb') as rpositionfile:
data = csv.reader(rpositionfile, skipinitialspace=True)
for row in data: real_position = row
rpositionfile.close()
print "Real position of the object is: ", real_position
point = geometric_transformation.geometric_transformation(0,0,0)
print dir(point)
# for every camera roto-traslate the object
with open(CAMERAS_DATA_FILE, 'rb') as csvfile:
data = csv.reader(csvfile, skipinitialspace=True)
for row in data:
# skip row with comments or empty
if len(row) == 0: continue
if row[0][0] == '#': continue
# transform text to integers
row = map(lambda x: float(x), row)
