-
Notifications
You must be signed in to change notification settings - Fork 1
/
datasets.py
139 lines (110 loc) · 3.65 KB
/
datasets.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
# -*- coding: utf-8 -*-
import numpy as np
import pandas as pd
import torch
from sklearn import preprocessing
import scipy.ndimage
from sklearn.decomposition import PCA
def wamp(data):
i = 0
N = len(data)
f_data = np.zeros(N-200)
while i < N-200:
tmp = data[i:i+199]-data[i+1:i+200]
f_data[i] = np.sum(tmp > 10)
i = i + 1
return f_data
def wl(data):
i = 0
N = len(data)
f_data = np.zeros(N-200)
while i < N-200:
tmp = np.abs(data[i:i+199]-data[i+1:i+200])
f_data[i] = np.sum(tmp)/200
f_data[i] = f_data[i]
i = i + 1
return f_data
def time_zc(data):
length = len(data)
diff = np.abs(data[0:length-1]-data[1:length])
mult = data[0:length-1]*data[1:length]
tmp = np.logical_and(diff>10, mult<0)
return np.sum(tmp)
def zc(data):
i = 0
N = len(data)
f_data = np.zeros(N-200)
while i < N-200:
f_data[i] = time_zc(data[i:i+200])
i = i + 1
return f_data
def mav(data):
i = 0
N = len(data)
f_data = np.zeros(N-200)
while i < N-200:
f_data[i] = np.mean(np.abs(data[i:i+200]))
i = i + 1
return f_data
#
def rms(data):
i = 0
N = len(data)
f_data = np.zeros(N-200)
while i < N-200:
f_data[i] = np.sqrt(np.mean(np.power(data[i:i+200], 2)))#/200
i = i + 1
return f_data
def data():
sEMG_sets = pd.read_csv('test_data/emg.csv')
sEMG = sEMG_sets['value'].to_numpy()
angle_sets = pd.read_csv('test_data/angle.txt', sep='\\s+')
AngleY = angle_sets['AngleY(deg)'].to_numpy()
AngleY = np.array(AngleY).ravel()
AngleY = scipy.ndimage.zoom(AngleY, 2, order=0)
y = np.expand_dims(AngleY, axis=-1)
y = preprocessing.MinMaxScaler().fit_transform(y)
return sEMG, y
def training_data():
sEMG, y = data()
f_wamp = wamp(sEMG)
f_wamp = np.expand_dims(f_wamp, axis=-1)
f_wamp = preprocessing.MinMaxScaler().fit_transform(f_wamp)
f_wamp = f_wamp.ravel()
f_mav = mav(sEMG)
f_mav = np.expand_dims(f_mav, axis=-1)
f_mav = preprocessing.MinMaxScaler().fit_transform(f_mav)
f_mav = f_mav.ravel()
f_rms = rms(sEMG)
f_rms = np.expand_dims(f_rms, axis=-1)
f_rms = preprocessing.MinMaxScaler().fit_transform(f_rms)
f_rms = f_rms.ravel()
f_wl = wl(sEMG)
f_wl = np.expand_dims(f_wl, axis=-1)
f_wl = preprocessing.MinMaxScaler().fit_transform(f_wl)
f_wl = f_wl.ravel()
f_zc = zc(sEMG)
f_zc = np.expand_dims(f_zc, axis=-1)
f_zc = preprocessing.MinMaxScaler().fit_transform(f_zc)
f_zc = f_zc.ravel()
feature = np.array([f_wamp, f_rms, f_mav, f_wl, f_zc]).T
pca = PCA(n_components=2)
new_feature = pca.fit_transform(feature)
feature_1 = new_feature[:,0]
feature_2 = new_feature[:,1]
X_feature1 = np.zeros((len(y)-4, 100))
X_feature2 = np.zeros((len(y)-4, 100))
y_tmp = np.zeros((len(y)-4, 1))
for i in range(len(y)-4):
if i == len(y)-5 and (len(y)-5) % 2 == 1:
y_tmp[i] = y[i]
X_feature1[i, 0:50] = feature_1[50 * i:50 * i + 50]
X_feature2[i, 0:50] = feature_2[50 * i:50 * i + 50]
else:
y_tmp[i] = y[i]
X_feature1[i, :] = feature_1[50*i:50*i+100]
X_feature2[i, :] = feature_2[50*i:50*i+100]
y = y_tmp
return torch.from_numpy(X_feature1).unsqueeze(dim=1).type(dtype='torch.FloatTensor'),\
torch.from_numpy(X_feature2).unsqueeze(dim=1).type(dtype='torch.FloatTensor'),\
torch.from_numpy(y).type(dtype='torch.FloatTensor')