A Hybrid CNN-GRU Approach for Enhanced LOS/NLOS Signal Classification

A Hybrid CNN-GRU Approach for Enhanced LOS/NLOS Signal Classification

연구 분야: Infrastructure

학회: 2024 9th International Conference on Communication and Electronics Systems (ICCES)

Wireless communication relies heavily on Line-of- Sight (LoS) and Non-Line-of-Sight (NLoS) classification, especially in situations with many barriers where precise prediction is difficult. To improve the classification of LoS and NLoS scenarios, we present a model in this study that combines Convolutional Neural Networks (CNNs) and Gated Recurrent Units (GRUs). Using their capacity to recognize complex patterns and localized correlations in high-dimensional multipath signal profiles, CNNs are used to extract spatial characteristics from signal data. By capturing temporal dependencies, GRUs make it possible to analyze sequential signal fluctuations more thoroughly. The first step in the suggested method is gathering signal data and generating multipath and delay-speed profiles. The accuracy of LoS/NLoS classification is then greatly increased by applying CNNs and GRUs to the data. This approach improves wireless communication performance under difficult circumstances by providing a reliable solution for settings with significant impediments.