Indoor Visible Light Fingerprint Positioning Scheme Based on Convolution Neural Network

Table 1. Fingerprint database
View table
Table 1. Fingerprint database
Reference point Coordinate Received signal strength
1 x₁ y₁ p₁₁ p₁₂ p₁₃ p₁₄
2 x₂ y₂ p₂₁ p₂₂ p₂₃ p₂₄
3 x₃ y₃ p₃₁ p₃₂ p₃₃ p₃₄
$⋮$ $⋮$ $⋮$ $⋮$ $⋮$ $⋮$ $⋮$
N x_N y_N p_N₁ p_N₂ p_N₃ p_N₄

Table 2. Parameters used in channel simulation

View table

Table 2. Parameters used in channel simulation

Parameter	Value	Parameter	Value
Transmitted optical power（P_t）/W	1	Absolute temperature（T_K）/K	300
Fied of view（FOV）/（°）	60	Dark current（I_bg）/μA	5100
Semi-angle at half power（φ_1/2）/（°）	70	Open-circuit voltage gain（G）	10
Refractive index	1.5	FET transconductance（g_m）/mS	30
Wall reflectivity（ρ）/%	80	Noise bandwidth factor（I₂）	0.562
Gain of optical filter	1	Noise bandwidth factor（I₃）	0.0868
Modulation bandwidth（B）/MHz	100	Noise factor（Γ）	1.5

Table 3. Comparison of positioning errors in different methods
View table
Table 3. Comparison of positioning errors in different methods
Location algorithm Max location error Average location error
CNN 25.02 10.13
Trilateral method 82.03 28.57
SVM 46.22 17.41
Decision tree 42.43 17.79

Tools

Get Citation

Copy Citation Text

Hao Xu, Xudong Wang, Nan Wu. Indoor Visible Light Fingerprint Positioning Scheme Based on Convolution Neural Network[J]. Laser & Optoelectronics Progress, 2021, 58(17): 1706008

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites