NN-LCS: Neural Network and Linear Coordinate Solver Fusion Method for UWB Localization in Car Keyless Entry System

Area of Science:

• Wireless communication
• Localization algorithms
• Embedded systems

Background:

• Ultra-wideband (UWB) technology offers precise localization for car keyless entry systems (KES).
• Non-line-of-sight (NLOS) conditions, caused by vehicle obstructions, introduce significant ranging errors in UWB systems.
• Existing methods to mitigate NLOS errors, such as neural networks, often suffer from low accuracy, overfitting, and large parameter counts.

Purpose of the Study:

• To propose a novel fusion method, the neural network and linear coordinate solver (NN-LCS), to enhance UWB localization accuracy in KES.
• To address the limitations of existing NLOS error mitigation techniques, including low accuracy and high computational complexity.
• To develop an end-to-end localization solution suitable for embedded deployment.

Main Methods:

• A fusion approach combining a neural network with a linear coordinate solver (NN-LCS) was developed.
• Two fully connected (FC) layers were used to extract distance and received signal strength (RSS) features.
• A multi-layer perceptron (MLP) integrated these features for distance estimation, utilizing least squares for error backpropagation.

Main Results:

• The NN-LCS method demonstrated high accuracy in UWB localization, effectively mitigating NLOS errors.
• The proposed model achieved a small size, making it suitable for deployment on embedded devices with limited computing power.
• The end-to-end approach directly outputs localization results, simplifying the system architecture.

Conclusions:

• The NN-LCS method provides a robust and accurate solution for UWB localization in car keyless entry systems, particularly under NLOS conditions.
• The efficiency and small model size of the NN-LCS facilitate its implementation in resource-constrained embedded systems.
• This approach offers a significant improvement over existing methods for precise and secure keyless entry localization.