def add_arrow(self, start, end):
self._pylab.gca().add_patch(
pylab.Arrow(start[0],
start[1],
end[0] - start[0],
end[1] - start[1],
width=0.1))
interpolation='nearest',
extent=(mx.min(), mx.max(), my.min(), my.max()),
cmap=pylab.cm.Blues,
aspect='auto',
origin='lower')
c2a, c2b, c2c = km.cluster_centers_
pylab.scatter(km.cluster_centers_[:, 0],
km.cluster_centers_[:, 1],
marker='x',
linewidth=2,
s=100,
color='black')
# import pdb;pdb.set_trace()
pylab.gca().add_patch(
pylab.Arrow(c1a[0], c1a[1], c2a[0] - c1a[0], c2a[1] - c1a[1], width=0.1))
pylab.gca().add_patch(
pylab.Arrow(c1b[0], c1b[1], c2b[0] - c1b[0], c2b[1] - c1b[1], width=0.1))
pylab.gca().add_patch(
pylab.Arrow(c1c[0], c1c[1], c2c[0] - c1c[0], c2c[1] - c1c[1], width=0.1))
pylab.savefig(os.path.join("..", "1400_03_0%i.png" % i))
pylab.clf()
i += 1
#################### 3 iterations ####################
km = KMeans(init='random',
n_clusters=num_clusters,
verbose=1,
n_init=1,
def kmeans_simple(F, P):
# N : Numero total de objetos
# M : Numero de caracteristicas
N, M = F.shape
# Critério de parada
m = np.Inf
m_min = 0.01
# Contador
i = 0
while m > m_min:
print '----------'
print 'Iteracao ', i,
# Calcula a matriz de distancias inicial (D):
D = distance.cdist(F, P, metric='euclidean')
print '\nMatriz de distancias(D):'
print D
# Objetos mais proximos do vetor 1
L1 = F[D[:, 0] <= D[:, 1]]
# Objetos mais proximos do vetor 2
L2 = F[D[:, 0] > D[:, 1]]
# --- Print ---
print '\nPontos em #1:'
print L1
print '\nPontos em #1:'
print L2
# Armazena valores dos vetores na ultima iteracao
P_old = P.copy()
# Move os vetores de prototipo
P = np.zeros([2, M])
P[0, :] = L1.mean()
P[1, :] = L2.mean()
# --- PRINT ---
print '\nP_old:'
print P_old
print '\nP: '
print P
# Distancia entre a posicao dos vetores atual e anterior.
m = np.max(np.sqrt(((P_old - P)**2).sum(0)))
print 'm: '
print m
# Pepara para plotar
class_C1 = L1
class_C2 = L2
title = 'K-means. Iteracao ' + str(i)
plt.figure()
plt.subplot(111)
plt.axis('equal')
plt.scatter(class_C1[:, 0], class_C1[:, 1], c='red')
plt.scatter(class_C2[:, 0], class_C2[:, 1], c='blue')
plt.scatter(P_old[0, 0], P_old[0, 1], marker='x', c='red')
plt.scatter(P_old[1, 0], P_old[1, 1], marker='x', c='blue')
ax = plt.gca()
ax.add_patch(
plt.Arrow(P_old[0, 0], P_old[0, 1], P[0, 0] - P_old[0, 0],
P[0, 1] - P_old[0, 1]))
ax.add_patch(
plt.Arrow(P_old[1, 0], P_old[1, 1], P[1, 0] - P_old[1, 0],
P[1, 1] - P_old[1, 1]))
plt.scatter(P[0, 0], P[0, 1], marker='x', c='red')
plt.scatter(P[1, 0], P[1, 1], marker='x', c='blue')
plt.title(title)
plt.xlabel('Caracteristica 0')
plt.ylabel('Caracteristica 1')
# Mostra as figuras na tela.
plt.show()
i = i + 1
# ---------------------------------------------------------------------
# Calcula a matriz de distancias inicial (D):
D = distance.cdist(F, P, metric='euclidean')
# Gera valores para o retorno
CLASS = np.zeros(N)
# Objetos mais proximos do vetor 1
CLASS[D[:, 0] <= D[:, 1]] = 0
# Objetos mais proximos do vetor 2
CLASS[D[:, 0] > D[:, 1]] = 1
# Pepara para plotar
class_C1 = F[CLASS == 0, :]
class_C2 = F[CLASS == 1, :]
plt.figure()
plt.subplot(111)
plt.axis('equal')
plt.scatter(class_C1[:, 0], class_C1[:, 1], c='red')
plt.scatter(class_C2[:, 0], class_C2[:, 1], c='blue')
plt.scatter(P[0, 0], P[0, 1], marker='x', c='red')
plt.scatter(P[1, 0], P[1, 1], marker='x', c='blue')
plt.title('K-means. Final.')
plt.xlabel('Caracteristica 0')
plt.ylabel('Caracteristica 1')
# Mostra as figuras na tela.
plt.show()
return [CLASS, P]
def visualization(robot, pr, factor=7):
plt.figure("Robot in the world", figsize=(field.w_width, field.w_length))
plt.title('Particle filter')
# draw coordinate grid for plotting
grid = [
-field.w_width / 2.0, field.w_width / 2.0, -field.w_length / 2.0,
field.w_length / 2.0
]
ax = plt.axis(grid)
landmarks = {
"blue_posts": [[-1.0, -4.5], [1.0, -4.5]],
"yellow_posts": [[-1.0, 4.5], [1.0, 4.5]]
}
for el in field.field:
if el == 'circles':
for circle in field.field['circles']:
plot_circle = plt.Circle((circle[0], circle[1]),
circle[2],
linewidth=2,
fill=False,
edgecolor='#330000')
plt.gca().add_patch(plot_circle)
if el == 'lines':
for line in field.field['lines']:
plot_line = plt.Line2D(line[0],
line[1],
linewidth=2,
linestyle="-",
color='#330000')
plt.gca().add_line(plot_line)
if el == 'rectangles':
for rectangle in field.field['rectangles']:
rect = plt.Rectangle(rectangle[0],
rectangle[1],
rectangle[2],
linewidth=2,
linestyle="-",
fill=False,
edgecolor='#330000')
plt.gca().add_patch(rect)
'''
# draw particles
for ind in range(len(p)):
# particle
circle = plt.Circle((p[ind][0].x, p[ind][0].y), 1./factor/2, facecolor='#ffb266', edgecolor='#994c00', alpha=0.5)
plt.gca().add_patch(circle)
# particle's orientation
arrow = plt.Arrow(p[ind][0].x, p[ind][0].y, 2*math.cos(p[ind][0].yaw)/factor, 2*math.sin(p[ind][0].yaw)/factor, width=1/factor, alpha=1., facecolor='#994c00', edgecolor='#994c00')
plt.gca().add_patch(arrow)
'''
# draw resampled particles
for ind in range(len(pr)):
# particle
circle = plt.Circle((pr[ind][0].y, pr[ind][0].x),
1. / factor / 2,
facecolor='#ffb266',
edgecolor='#cc0000',
alpha=0.5)
plt.gca().add_patch(circle)
# particle's orientation
arrow = plt.Arrow(pr[ind][0].y,
pr[ind][0].x,
2 * math.sin(pr[ind][0].yaw) / factor,
math.cos(pr[ind][0].yaw) / factor,
width=1 / factor,
alpha=1.,
facecolor='#006600',
edgecolor='#006600')
plt.gca().add_patch(arrow)
# robot's location
circle = plt.Circle((robot.y, robot.x),
1. / factor,
facecolor='#FF66E9',
edgecolor='#FF66E9')
plt.gca().add_patch(circle)
# robot's orientation
arrow = plt.Arrow(robot.y,
robot.x,
3 * math.sin(robot.yaw) / factor,
3 * math.cos(robot.yaw) / factor,
width=1.0 / factor,
alpha=0.5,
facecolor='#000000',
edgecolor='#000000')
plt.gca().add_patch(arrow)
#fixed landmarks of known locations2
for lm in landmarks:
for lms in landmarks[lm]:
if lm == "yellow_posts":
circle = plt.Circle(((lms[0], lms[1])),
1. / factor,
facecolor='#ffff00',
edgecolor='#330000')
plt.gca().add_patch(circle)
else:
circle = plt.Circle(((lms[0], lms[1])),
1. / factor,
facecolor='#060C73',
edgecolor='#330000')
plt.gca().add_patch(circle)
