Efficient coding and detection of ultra-long IDs for visible light positioning systems

Area of Science:

• Computer Vision
• Wireless Communication
• Indoor Localization

Background:

• Visible Light Positioning (VLP) complements Global Navigation Satellite Systems (GNSS) for indoor environments where satellite signals are weak.
• Large-scale VLP deployment requires efficient detection of numerous Light Emitting Diode (LED) IDs, posing challenges for image sensors due to long IDs spanning multiple frames.
• Rolling shutter cameras are commonly used but face difficulties in capturing complete, ultra-long LED IDs.

Purpose of the Study:

• To investigate and propose a novel coding technique for detecting ultra-long LED IDs using rolling shutter cameras.
• To enhance the efficiency and accuracy of LED ID detection in Visible Light Positioning systems.
• To develop a coding method that balances performance with reduced decoding complexity.

Main Methods:

• Analysis of data loss patterns in rolling shutter camera frames.
• Development of a novel coding technique tailored for LED ID detection.
• Performance evaluation of Reed-Solomon (RS) codes and a proposed new coding method.
• Numerical simulations and experimental validation.

Main Results:

• The proposed coding technique significantly decreases the number of frames required for data processing.
• Detection time for ultra-long LED IDs is substantially reduced.
• The accuracy of LED ID detection is improved.
• The novel coding method achieves performance comparable to RS codes with considerably lower decoding complexity.

Conclusions:

• The developed coding technique effectively enables the detection of much longer LED IDs.
• This advancement is crucial for the large-scale implementation of Visible Light Positioning for global indoor localization.
• The proposed method offers a practical solution by optimizing the trade-off between detection performance and computational cost.