High-Precision Indoor Visible Light Positioning Using Modified Momentum Back Propagation Neural Network with Sparse...

Area of Science:

• Electrical Engineering
• Computer Science
• Robotics

Background:

• Accurate indoor positioning is crucial for various applications.
• Existing visible light positioning systems often require extensive training data.
• Received Signal Strength (RSS) based methods can be sensitive to data sparsity.

Purpose of the Study:

• To propose and evaluate a Modified Momentum Back-Propagation (MMBP) algorithm for indoor visible light positioning.
• To demonstrate high-precision localization using a sparse training dataset.
• To compare the proposed algorithm's performance against traditional RSS algorithms.

Main Methods:

• Developed an indoor visible light positioning technique utilizing the MMBP algorithm.
• Employed a sparse training dataset (20 points) for localization within a defined area (1.8 m × 1.8 m × 2.1 m).
• Experimentally validated the MMBP algorithm with both even and arbitrary training data acquisition methods.

Main Results:

• Achieved average localization accuracies of 1.88 cm (arbitrary set) and 1.99 cm (even set).
• Demonstrated a 7.6 times higher positioning accuracy compared to the traditional RSS algorithm (14.34 cm error).
• Outperformed previous RSS-based and complex machine learning algorithms requiring large datasets.

Conclusions:

• The MMBP algorithm enables high-precision indoor visible light positioning with significantly reduced training data.
• The proposed method offers a robust and efficient solution for indoor localization challenges.
• This technique presents a promising advancement for visible light communication and positioning systems.