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
| import numpy as np
import torch
import torch.nn as nn
from typing import Dict, Tuple
class CabinThermalVirtualSensor(nn.Module):
"""
座舱热舒适虚拟传感器
基于 MDPI 2025 论文：
"Development of Virtual Sensors for Vehicle In-Cabin
Temperature Prediction Using Deep Learning"
用少量物理传感器 + AI 推断全舱温度分布
物理传感器（6个）：
| 位置 | 传感器类型 | 参数 |
|------|-----------|------|
| 仪表盘 | 温度+湿度 | ±0.5°C, ±3%RH |
| 车顶 | 温度 | ±0.5°C |
| 副驾座椅 | 温度 | ±0.5°C |
| 后排左 | 温度 | ±0.5°C |
| 后排右 | 温度 | ±0.5°C |
| 挡风玻璃内侧 | 温度 | ±0.5°C |
虚拟推断点（20+）：
- 驾驶员面部温度
- 驾驶员手部温度
- 方向盘温度
- 座椅表面温度
- 脚部空间温度
- 后视镜附近温度
"""
def __init__(self, num_physical: int = 6, num_virtual: int = 20):
super().__init__()
self.encoder = nn.Sequential(
nn.Linear(num_physical + 4, 64),
nn.ReLU(),
nn.Linear(64, 128),
nn.ReLU(),
nn.Linear(128, 64)
)
self.decoder = nn.Sequential(
nn.Linear(64, 128),
nn.ReLU(),
nn.Linear(128, 64),
nn.ReLU(),
nn.Linear(64, num_virtual)
)
self.pmv_head = nn.Sequential(
nn.Linear(num_virtual + 2, 32),
nn.ReLU(),
nn.Linear(32, 1)
)
def forward(self, physical_sensors: torch.Tensor,
env_data: torch.Tensor) -> Dict[str, torch.Tensor]:
"""
Args:
physical_sensors: (B, 6) 物理传感器读数
env_data: (B, 4) [外温, 日照W/m², 车速, 风扇转速]
Returns:
output: {
virtual_temps: (B, 20) 虚拟温度点,
pmv: (B, 1) 预测平均投票
}
"""
x = torch.cat([physical_sensors, env_data], dim=-1)
latent = self.encoder(x)
virtual_temps = self.decoder(latent)
humidity = env_data[:, 0:1] * 0
air_speed = env_data[:, 2:3] * 0.01
pmv_input = torch.cat([virtual_temps, humidity, air_speed], dim=-1)
pmv = self.pmv_head(pmv_input)
return {
"virtual_temps": virtual_temps,
"pmv": pmv
}
def comfort_level(self, pmv: float) -> str:
"""PMV → 舒适等级"""
if -0.5 <= pmv <= 0.5:
return "comfortable"
elif -1.0 <= pmv < -0.5 or 0.5 < pmv <= 1.0:
return "slightly_uncomfortable"
elif -1.5 <= pmv < -1.0 or 1.0 < pmv <= 1.5:
return "uncomfortable"
else:
return "very_uncomfortable"
class ThermalDMSIntegration:
"""
热舒适 + DMS 融合：温度影响疲劳的联合分析
研究发现：
- 高温（>28°C）加速 PERCLOS 增长
- 低温（<18°C）增加眨眼频率
- 温度波动导致注意力分散
- 湿度>70%加速主观疲劳感
融合架构：
DMS PERCLOS + 热舒适 PMV → 综合疲劳风险评估
"""
TEMP_PERCLOS_FACTOR = {
(18, 22): 1.0,
(22, 26): 1.0,
(26, 28): 1.2,
(28, 30): 1.5,
(30, 32): 1.8,
(32, 35): 2.5,
}
def adjust_fatigue_score(self, perclos: float, cabin_temp: float,
humidity: float) -> dict:
"""
根据热环境调整疲劳分数
Args:
perclos: DMS 检测的 PERCLOS 值 (%)
cabin_temp: 座舱温度 (°C)
humidity: 相对湿度 (%)
Returns:
adjusted: 调整后的疲劳评估
"""
temp_factor = 1.0
for (lo, hi), factor in self.TEMP_PERCLOS_FACTOR.items():
if lo <= cabin_temp < hi:
temp_factor = factor
break
if humidity > 70:
humidity_factor = 1.3
elif humidity > 60:
humidity_factor = 1.1
else:
humidity_factor = 1.0
adjusted_perclos = perclos * temp_factor * humidity_factor
return {
"raw_perclos": perclos,
"adjusted_perclos": round(adjusted_perclos, 1),
"temp_factor": temp_factor,
"humidity_factor": humidity_factor,
"thermal_alert": cabin_temp > 28 or cabin_temp < 18,
}
|
