def draw():
data=md.loadData('result.tl')
for n,(k,v) in enumerate(data.iteritems()):
mat=v[2]['matrix']
path='fig/fig_%s'%k
os.mkdir(path)
for i in range(0,len(mat),1):
drawHeatMap(mat[i],os.path.join(path,'%d.png'%i),3*k,i*div[n])
def draw():
data = md.loadData('result.tl')
for n, (k, v) in enumerate(data.iteritems()):
mat = v[2]['matrix']
path = 'fig/fig_%s' % k
os.mkdir(path)
for i in range(0, len(mat), 1):
drawHeatMap(mat[i], os.path.join(path, '%d.png' % i), 3 * k,
i * div[n])
def draw1():
data=md.loadData('result1.tl')
n=4
k=25
v=data[k]
mat=v[3]['matrix']
path='fig/fig_%s'%k
if not os.path.exists(path):
os.mkdir(path)
for i in range(0,len(mat),2):
drawHeatMap(mat[i],os.path.join(path,'%d.png'%int(i/2)),3*k,i*div[n])
def draw1():
data = md.loadData('result1.tl')
n = 4
k = 25
v = data[k]
mat = v[3]['matrix']
path = 'fig/fig_%s' % k
if not os.path.exists(path):
os.mkdir(path)
for i in range(0, len(mat), 2):
drawHeatMap(mat[i], os.path.join(path, '%d.png' % int(i / 2)), 3 * k,
i * div[n])
def draw():
global SL
data=md.loadData('result.tl')
for k,v in data.iteritems():
SL=14.0/k
L=[]
for i in range(k-1):
L.append(-7+(i+1)*SL-0.01)
X1=np.insert(X,0,L)
X1.sort()
plt.figure(k)
plt.plot(X,Y,'b.',alpha=0.6,label="measure point")
plt.plot(X1,map(realLineFunc(v[0]['param']),X1),c="red",lw=2,ls="-",alpha=0.7,label="M1")
plt.plot(X1,map(realLineFunc(v[1]['param']),X1),c="blue",lw=2,ls="-",alpha=0.7,label="M2")
plt.plot(X1,map(realLineFunc(v[2]['param']),X1),c="black",lw=2,ls="--",alpha=0.7,label="M3")
plt.plot(X1,map(lambda x:10*math.sin(0.6*x),X1),c="cyan",lw=2,ls="-",alpha=0.7,label="real function")
plt.legend(loc='best')
plt.title("fitting interval sin curve with quadratic curve by %s partition"%k)
plt.xlabel('x')
plt.ylabel('y')
plt.savefig('img/img_%s.pdf'%k)
__author__ = "luzhijun"
'''
cma restart test
'''
import cma
import math
import numpy as np
import matplotlib.pyplot as plt
from multiprocessing import Pool
import makeData as md
import time
plt.rc('figure', figsize=(16, 9))
PI = math.pi
E = math.exp
data = md.loadData('data.tl')
X = data[0]
Y = data[1]
RANGE = max(Y) - min(Y)
PN = 300
ALPHA = 0.01
BETA = PI / 18.0
PARTITION = 5
DIM = 3 * PARTITION
SPN = int(PN / PARTITION)
SL = 14.0 / PARTITION
L = []
sys.path.append('..')
import makeData
shuffling = True
scaling = True
max_cmd_vel_linear_x = 0.5
max_cmd_vel_angular_z = 1
bias_cmd_vel_angular = 0
max_iteration = 10001
disp_epoch = 50
save_epoch = 2000
#imgwidth=80;
#imgheight=60;
# first make data
inputs, labels = makeData.loadData('../data/data518/', resize=None)
batchsize = 16
Ndata = len(labels)
vmode = 0 #0 for speed, 1 for angular velocity
images = []
cmdvels = []
randomidx = range(Ndata)
random.shuffle(randomidx)
for i in range(Ndata):
images.append(inputs[randomidx[i]])
cmdvels.append(labels[randomidx[i]])
#!usr/bin/env python
#encoding: utf-8
__author__="luzhijun"
'''
make linear data sets
'''
import numpy as np
import math
import makeData as md
import matplotlib.pyplot as plt
plt.rc('figure', figsize=(16, 9))
dataSet=md.loadData('data.tl')
X=dataSet[0]
Y=dataSet[1]
def realLineFunc(param):
def f(x):
i=int(math.floor((x+7)/SL))
return np.poly1d(param[3*i:3*(i+1)])(x)
return f
def draw():
global SL
data=md.loadData('result.tl')
for k,v in data.iteritems():
SL=14.0/k
L=[]
__author__="luzhijun"
'''
cma restart test
'''
import cma
import math
import numpy as np
import matplotlib.pyplot as plt
from multiprocessing import Pool
import makeData as md
import time
plt.rc('figure', figsize=(16, 9))
PI=math.pi
E=math.exp
data=md.loadData('data.tl')
X=data[0]
Y=data[1]
RANGE=max(Y)-min(Y)
PN=300
ALPHA=0.01
BETA=PI/18.0
PARTITION=5
DIM=3*PARTITION
SPN=int(PN/PARTITION)
SL=14.0/PARTITION
L=[]
