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
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
| """
风险预测 HMI 评估管道
基于 Liang et al., 2026 实验设计
功能:
1. 模拟 3×2×2 实验设计
2. 生成 fNIRS/ECG/行为数据
3. 统计分析 HMI 效果
4. 可视化结果
依赖:
pip install numpy scipy scikit-learn matplotlib
"""
import numpy as np
from scipy import stats
from dataclasses import dataclass
from typing import List, Dict, Tuple
from enum import IntEnum
import json
class HMIType(IntEnum):
BASELINE = 0
RISK_ALERT = 1
MULTIMODAL = 2
class RiskLevel(IntEnum):
LOW = 0
HIGH = 1
@dataclass
class TrialResult:
"""单次试验结果"""
hmi: HMIType
risk: RiskLevel
turn: str
takeover: bool
takeover_time: float
subjective_risk: float
situational_trust: float
workload: float
fnirs_left_ba10: float
fnirs_right_ba10: float
hrv_sdnn: float
eda_peak: float
class RiskPredictiveHMISimulator:
"""风险预测 HMI 效果仿真器"""
def __init__(self, n_participants: int = 30):
self.n = n_participants
np.random.seed(42)
def generate_trial(self, p_id: int, hmi: HMIType,
risk: RiskLevel, turn: str) -> TrialResult:
"""生成单次试验数据"""
p_baseline = np.random.normal(0, 0.1)
hmi_effects = {
HMIType.BASELINE: {
'risk_diff': 1.1, 'trust': 5.5, 'takeover_rate': 0.65,
'fnirs_left': 0.35, 'fnirs_right': 0.28
},
HMIType.RISK_ALERT: {
'risk_diff': 2.5, 'trust': 6.0, 'takeover_rate': 0.45,
'fnirs_left': 0.52, 'fnirs_right': 0.40
},
HMIType.MULTIMODAL: {
'risk_diff': 3.5, 'trust': 7.2, 'takeover_rate': 0.35,
'fnirs_left': 0.68, 'fnirs_right': 0.55
}
}
eff = hmi_effects[hmi]
risk_level = 8.0 if risk == RiskLevel.HIGH else 4.5
subj_risk = risk_level + np.random.normal(0, 1.0) - (eff['risk_diff'] - 1.5) * 0.3
trust = eff['trust'] + np.random.normal(0, 0.8)
if risk == RiskLevel.HIGH:
trust -= 1.5
if turn == 'right':
trust -= 0.5
if risk == RiskLevel.HIGH:
takeover_prob = eff['takeover_rate'] + 0.2
else:
takeover_prob = eff['takeover_rate'] - 0.2
takeover = np.random.random() < takeover_prob
takeover_time = np.random.exponential(2.5) if takeover else 0.0
workload = 5.0 + (1 if risk == RiskLevel.HIGH else 0) * 2 + \
(1 if hmi == HMIType.MULTIMODAL else 0) * 0.5 + \
np.random.normal(0, 0.8)
fnirs_left = eff['fnirs_left'] + p_baseline + \
(0.2 if risk == RiskLevel.HIGH else 0) + \
np.random.normal(0, 0.08)
fnirs_right = eff['fnirs_right'] + p_baseline + \
(0.15 if risk == RiskLevel.HIGH else 0) + \
np.random.normal(0, 0.08)
hrv = 45 + (5 if hmi == HMIType.MULTIMODAL else 0) + \
(-10 if risk == RiskLevel.HIGH else 0) + \
np.random.normal(0, 5)
eda = 2.0 + (1.5 if risk == RiskLevel.HIGH else 0) + \
np.random.normal(0, 0.5)
return TrialResult(
hmi=hmi, risk=risk, turn=turn,
takeover=takeover, takeover_time=takeover_time,
subjective_risk=np.clip(subj_risk, 1, 10),
situational_trust=np.clip(trust, 1, 10),
workload=np.clip(workload, 1, 10),
fnirs_left_ba10=max(0, fnirs_left),
fnirs_right_ba10=max(0, fnirs_right),
hrv_sdnn=max(20, hrv),
eda_peak=max(0.5, eda)
)
def run_experiment(self) -> List[TrialResult]:
"""运行完整实验"""
results = []
for p_id in range(self.n):
for hmi in HMIType:
for risk in RiskLevel:
for turn in ['left', 'right']:
results.append(self.generate_trial(p_id, hmi, risk, turn))
return results
def analyze(self, results: List[TrialResult]) -> Dict:
"""统计分析"""
analysis = {}
for hmi in HMIType:
hmi_results = [r for r in results if r.hmi == hmi]
high = [r for r in hmi_results if r.risk == RiskLevel.HIGH]
low = [r for r in hmi_results if r.risk == RiskLevel.LOW]
analysis[hmi.name] = {
'n': len(hmi_results),
'takeover_rate_high': np.mean([r.takeover for r in high]),
'takeover_rate_low': np.mean([r.takeover for r in low]),
'subj_risk_high': np.mean([r.subjective_risk for r in high]),
'subj_risk_low': np.mean([r.subjective_risk for r in low]),
'trust': np.mean([r.situational_trust for r in hmi_results]),
'workload': np.mean([r.workload for r in hmi_results]),
'fnirs_left': np.mean([r.fnirs_left_ba10 for r in high]),
'fnirs_right': np.mean([r.fnirs_right_ba10 for r in high]),
'hrv_high': np.mean([r.hrv_sdnn for r in high]),
'hrv_low': np.mean([r.hrv_sdnn for r in low]),
}
risk_high = [r.subjective_risk for r in high]
risk_low = [r.subjective_risk for r in low]
t_stat, p_val = stats.ttest_ind(risk_high, risk_low)
analysis[hmi.name]['risk_t_stat'] = round(t_stat, 2)
analysis[hmi.name]['risk_p_value'] = round(p_val, 4)
return analysis
if __name__ == "__main__":
print("=" * 70)
print("风险预测 HMI 评估管道")
print("基于 Liang et al., 2026")
print("=" * 70)
sim = RiskPredictiveHMISimulator(n_participants=30)
results = sim.run_experiment()
print(f"\n实验完成: {len(results)} 次试验 (30被试 × 3 HMI × 2 风险 × 2 转向)")
analysis = sim.analyze(results)
print(f"\n{'='*70}")
print("HMI 效果对比")
print(f"{'='*70}")
print(f"\n{'指标':<20s} {'Baseline':>10s} {'Risk-alert':>10s} {'Multimodal':>10s}")
print("-" * 55)
metrics = [
('高风险接管率', 'takeover_rate_high', '%'),
('低风险接管率', 'takeover_rate_low', '%'),
('高风险主观评分', 'subj_risk_high', '/10'),
('低风险主观评分', 'subj_risk_low', '/10'),
('情境信任', 'trust', '/10'),
('工作量', 'workload', '/10'),
('左前额BA10', 'fnirs_left', 'μM'),
('右前额BA10', 'fnirs_right', 'μM'),
('HRV (高风险)', 'hrv_high', 'ms'),
('HRV (低风险)', 'hrv_low', 'ms'),
]
for name, key, unit in metrics:
vals = []
for hmi in ['BASELINE', 'RISK_ALERT', 'MULTIMODAL']:
v = analysis[hmi][key]
if unit == '%':
vals.append(f"{v*100:.0f}%")
elif unit == '/10':
vals.append(f"{v:.1f}")
else:
vals.append(f"{v:.2f}")
print(f"{name:<20s} {vals[0]:>10s} {vals[1]:>10s} {vals[2]:>10s}")
print(f"\n{'='*70}")
print("高 vs 低风险区分能力 (t检验)")
print(f"{'='*70}")
print(f"{'HMI类型':<15s} {'t统计量':>10s} {'p值':>10s} {'显著性':>10s}")
print("-" * 50)
for hmi_name in ['BASELINE', 'RISK_ALERT', 'MULTIMODAL']:
t = analysis[hmi_name]['risk_t_stat']
p = analysis[hmi_name]['risk_p_value']
sig = '***' if p < 0.001 else ('**' if p < 0.01 else ('*' if p < 0.05 else 'ns'))
print(f"{hmi_name:<15s} {t:>10.2f} {p:>10.4f} {sig:>10s}")
print(f"\n{'='*70}")
print("IMS 应用启示")
print(f"{'='*70}")
print("""
1. 风险预测 HMI 设计:
- 视觉提示: HUD 显示风险等级颜色 (绿→黄→红)
- 听觉提示: 高风险时增加定向声音警报
- 多模态 > 仅视觉 > 无提示
2. 接管决策支持:
- 低风险: 抑制接管提示 (降低不必要接管)
- 高风险: 强化接管提示 (确保及时接管)
- 效果: 总接管次数 -48%, 低风险接管 -64%
3. AR-HUD 自适应:
┌─────────────────────────────────────────┐
│ 认知状态 AR-HUD 策略 │
│ ─────────────────────────────────── │
│ 低风险+清醒 标准导航+路况信息 │
│ 低风险+疲劳 增加刺激信息密度 │
│ 高风险+清醒 简化显示, 突出风险 │
│ 高风险+过载 仅显示关键警告+声音 │
└─────────────────────────────────────────┘
4. 部署路线:
Phase 1 (2026): 视觉风险提示 (HUD颜色编码)
Phase 2 (2027): +听觉多模态 (定向声音)
Phase 3 (2028): +fNIRS认知状态自适应
Phase 4 (2029): 完全自适应风险预测 HMI
""")
|
