Stability Optimization of Visible Light Indoor Positioning with Single LED: Using Elastic Weight Consolidation Mechanism Convolutional Neural Network

Le résumé fourni par la source

This paper proposes a continual learning strategy that combines Convolutional Neural Networks (CNNs) with the Elastic Weight Consolidation (EWC) mechanism. The objective is to optimize a single LED-based Visible Light Positioning (VLP) system, enhancing its stability and generalization in dynamic environments. The innovation lies in the weighted feedback mechanism of EWC, which enables the neural network to retain previously learned knowledge while acquiring new tasks, thereby improving generalization and reducing the need for extensive training data. Laboratory experiments demonstrated that the EWC-enhanced ResNet50 model maintained a positioning accuracy within 4 cm at a height of 160 cm, even under scenarios with human movement and sunlight interference, requiring only 27% of the data from the new environment to achieve a positioning accuracy of 5 cm.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.