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
| import numpy as np
from typing import Tuple
from scipy.signal import welch, find_peaks
def extract_csi_features(
csi_data: np.ndarray,
fs: float = 100.0,
motion_band: Tuple[float, float] = (0.5, 3.0)
) -> dict:
"""
从CSI数据中提取运动特征
论文核心算法复现：自适应PCA主成分选择
Args:
csi_data: CSI矩阵, shape=(n_subcarriers, n_samples)
fs: 采样率(Hz)
motion_band: 人体运动频段(Hz)
- 0.5-3 Hz: 呼吸/心跳微动
- 3-10 Hz: 肢体运动
Returns:
包含选择的主成分和检测结果的字典
参考:
Naotunna et al., "Through-Wall Detection using SDR
based on adaptive PCA", arXiv:2609.12443, 2026
"""
n_subcarriers, n_samples = csi_data.shape
from sklearn.decomposition import PCA
pca = PCA(n_components=min(10, n_subcarriers))
components = pca.fit_transform(csi_data.T)
scores = []
for i in range(components.shape[1]):
freqs, psd = welch(
components[:, i],
fs=fs,
nperseg=min(256, len(components[:, i]))
)
motion_mask = (freqs >= motion_band[0]) & (freqs <= motion_band[1])
motion_energy = np.sum(psd[motion_mask])
total_energy = np.sum(psd)
if total_energy > 0:
spectral_score = motion_energy / total_energy
else:
spectral_score = 0
noise_band = (freqs >= 10) & (freqs <= 20)
noise_power = np.mean(psd[noise_band]) if np.any(noise_band) else 1e-10
snr = motion_energy / noise_power
scores.append({
'component_idx': i,
'spectral_score': float(spectral_score),
'snr': float(snr),
'motion_energy': float(motion_energy),
'explained_variance': float(pca.explained_variance_ratio_[i])
})
scores.sort(key=lambda x: x['spectral_score'], reverse=True)
best_component_idx = scores[0]['component_idx']
best_signal = components[:, best_component_idx]
import pywt
scales = np.arange(1, 64)
wavelet = 'morl'
coeffs, freqs_cwt = pywt.cwt(best_signal, scales, wavelet, 1.0/fs)
energy = np.abs(coeffs)
energy_sum = np.sum(energy, axis=0)
threshold = np.mean(energy_sum) + 2 * np.std(energy_sum)
peaks, _ = find_peaks(energy_sum, height=threshold, distance=int(fs*0.5))
return {
'motion_detected': len(peaks) > 0,
'n_events': len(peaks),
'event_times': peaks / fs,
'best_component': best_component_idx,
'spectral_scores': scores,
'energy_profile': energy_sum.tolist(),
'detection_threshold': float(threshold)
}
def simulate_cabin_csi(
n_subcarriers: int = 56,
duration_s: float = 10.0,
fs: float = 100.0,
breathing_rate: float = 0.3,
motion_events: list = None
) -> np.ndarray:
"""
模拟座舱内CSI数据（含人体微动）
Args:
n_subcarriers: WiFi子载波数
duration_s: 持续时间(秒)
fs: 采样率(Hz)
breathing_rate: 呼吸频率(Hz), 婴儿~0.3-1.0
motion_events: 运动事件列表 [(start_s, duration_s, amplitude)]
Returns:
csi_data: shape=(n_subcarriers, n_samples)
"""
n_samples = int(duration_s * fs)
t = np.linspace(0, duration_s, n_samples)
if motion_events is None:
motion_events = [(2.0, 3.0, 0.5)]
static_csi = np.exp(1j * 2 * np.pi * 0.1 * np.random.randn(n_subcarriers))
csi = np.zeros((n_subcarriers, n_samples), dtype=complex)
for k in range(n_subcarriers):
csi[k, :] = static_csi[k]
breathing = 0.01 * np.sin(2 * np.pi * breathing_rate * t)
csi[k, :] += breathing * np.exp(-1j * 2 * np.pi * k * 0.01)
for start, dur, amp in motion_events:
mask = (t >= start) & (t <= start + dur)
motion = amp * np.random.randn(np.sum(mask))
csi[k, mask] += motion * np.exp(-1j * 2 * np.pi * k * 0.05)
noise = 0.01 * (np.random.randn(*csi.shape) + 1j * np.random.randn(*csi.shape))
csi += noise
return csi
if __name__ == "__main__":
print("=== WiFi CSI 穿墙/座舱感知检测仿真 ===\n")
np.random.seed(42)
print("[场景1] 后排有儿童（呼吸频率0.4Hz）")
csi_with_person = simulate_cabin_csi(
n_subcarriers=56,
duration_s=10.0,
fs=100.0,
breathing_rate=0.4,
motion_events=[(3.0, 1.0, 0.3), (6.5, 0.5, 0.2)]
)
result1 = extract_csi_features(csi_with_person, fs=100.0, motion_band=(0.3, 3.0))
print(f" 运动检测: {result1['motion_detected']}")
print(f" 检测到 {result1['n_events']} 个运动事件")
print(f" 最佳主成分: PC{result1['best_component']}")
print(f" 频谱评分: {result1['spectral_scores'][0]['spectral_score']:.4f}")
print(f" 事件时刻: {[f'{t:.2f}s' for t in result1['event_times']]}")
print("\n[场景2] 空座（无人体微动）")
csi_empty = simulate_cabin_csi(
n_subcarriers=56,
duration_s=10.0,
fs=100.0,
breathing_rate=0.0,
motion_events=[]
)
result2 = extract_csi_features(csi_empty, fs=100.0, motion_band=(0.3, 3.0))
print(f" 运动检测: {result2['motion_detected']}")
print(f" 检测到 {result2['n_events']} 个运动事件")
print(f" 最佳主成分: PC{result2['best_component']}")
print(f" 频谱评分: {result2['spectral_scores'][0]['spectral_score']:.4f}")
|
