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
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
| """
酒精损伤多分类器
"""
import numpy as np
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import LeaveOneOut
from sklearn.metrics import classification_report, roc_auc_score
from sklearn.preprocessing import StandardScaler
class AlcoholImpairmentDetector:
"""
基于运动学特征的酒精损伤检测器
特征：每个信号的排列熵 + 标准差
分类：清醒 / 轻度损伤 / 重度损伤
"""
def __init__(self, window_sec: float = 30.0, fps: int = 100):
self.window_sec = window_sec
self.fps = fps
self.window_samples = int(window_sec * fps)
self.signals = [
'imu_ax', 'imu_ay', 'imu_az',
'imu_gx', 'imu_gy', 'imu_gz',
'throttle'
]
self.classifier = LogisticRegression(
max_iter=1000,
multi_class='multinomial',
solver='lbfgs',
C=1.0
)
self.scaler = StandardScaler()
def extract_features(self, imu_data: np.ndarray,
throttle: np.ndarray) -> np.ndarray:
"""
从窗口数据中提取特征
Args:
imu_data: (window, 6) IMU数据 [ax, ay, az, gx, gy, gz]
throttle: (window,) 油门位置
Returns:
features: (14,) [PE×7, Std×7]
"""
features = []
signals = [
imu_data[:, 0],
imu_data[:, 1],
imu_data[:, 2],
imu_data[:, 3],
imu_data[:, 4],
imu_data[:, 5],
throttle
]
for sig in signals:
pe = permutation_entropy(sig, order=3, delay=1)
std = np.std(sig)
features.extend([pe, std])
return np.array(features)
def fit(self, X: np.ndarray, y: np.ndarray):
"""训练分类器"""
X_scaled = self.scaler.fit_transform(X)
self.classifier.fit(X_scaled, y)
def predict(self, X: np.ndarray) -> np.ndarray:
"""预测"""
X_scaled = self.scaler.transform(X)
return self.classifier.predict(X_scaled)
def predict_proba(self, X: np.ndarray) -> np.ndarray:
"""预测概率"""
X_scaled = self.scaler.transform(X)
return self.classifier.predict_proba(X_scaled)
def evaluate_loocv(self, X: np.ndarray, y: np.ndarray) -> dict:
"""
Leave-One-Participant-Out 评估
"""
from sklearn.model_selection import LeaveOneOut
loo = LeaveOneOut()
predictions = []
probabilities = []
for train_idx, test_idx in loo.split(X):
X_train, X_test = X[train_idx], X[test_idx]
y_train, y_test = y[train_idx], y[test_idx]
scaler = StandardScaler()
X_train_s = scaler.fit_transform(X_train)
X_test_s = scaler.transform(X_test)
clf = LogisticRegression(max_iter=1000, multi_class='multinomial')
clf.fit(X_train_s, y_train)
predictions.extend(clf.predict(X_test_s))
probabilities.extend(clf.predict_proba(X_test_s))
predictions = np.array(predictions)
probabilities = np.array(probabilities)
accuracy = np.mean(predictions == y)
from sklearn.preprocessing import label_binarize
y_bin = label_binarize(y, classes=[0, 1, 2])
auroc = roc_auc_score(y_bin, probabilities,
multi_class='ovr', average='weighted')
return {
'accuracy': accuracy,
'auroc': auroc,
'predictions': predictions,
'report': classification_report(y, predictions, output_dict=True)
}
class DrivingScenarioMigration:
"""
将e-scooter方法迁移到驾驶场景
关键映射：
e-scooter IMU → 车辆 IMU（或 CAN 总线加速度数据）
e-scooter 油门 → 车辆油门踏板位置
e-scooter 刹车杆 → 车辆刹车踏板位置
新增：方向盘转角（e-scooter没有）
"""
@staticmethod
def migrate_features():
"""e-scooter → 驾驶场景特征映射"""
mapping = {
'IMU_X 加速度': '车辆纵向加速度（CAN总线）',
'IMU_Y 加速度': '车辆横向加速度（CAN总线）',
'IMU_Z 加速度': '车辆垂直加速度（IMU）',
'油门位置': '油门踏板位置（CAN总线）',
'方向盘转角': '方向盘转角传感器（CAN总线）',
'方向盘转角速度': '方向盘转角变化率',
'刹车踏板': '刹车踏板位置（CAN总线）',
'车速': '车速（CAN总线）',
'IMU 陀螺仪': '车辆陀螺仪（可选，多数车无）'
}
return mapping
@staticmethod
def driving_feature_extraction(can_data: dict,
window_sec: float = 30.0,
fps: int = 100) -> np.ndarray:
"""
从CAN总线数据提取运动学特征
Args:
can_data: {
'steering_angle': (n,) 方向盘转角 (度),
'throttle': (n,) 油门踏板位置 (0-100%),
'brake': (n,) 刹车踏板位置 (0-100%),
'speed': (n,) 车速 (km/h),
'lateral_accel': (n,) 横向加速度 (m/s²),
'longitudinal_accel': (n,) 纵向加速度 (m/s²)
}
"""
window = int(window_sec * fps)
features = []
signals = [
can_data['steering_angle'],
can_data['throttle'],
can_data['brake'],
can_data['speed'],
can_data['lateral_accel'],
can_data['longitudinal_accel'],
]
for sig in signals:
pe = permutation_entropy(sig, order=3, delay=1)
std = np.std(sig)
features.extend([pe, std])
steering_rate = np.diff(can_data['steering_angle']) * fps
pe_rate = permutation_entropy(steering_rate, order=3, delay=1)
std_rate = np.std(steering_rate)
features.extend([pe_rate, std_rate])
return np.array(features)
|