"LSTM for UWB Positiong "

"Ultra-Wideband (UWB) positioning technology has been widely applied in indoor positioning, but its deployment is still hindered by non-ideal environmental factors such as Non-Line-of-Sight (NLOS) conditions and coplanar anchor placement, making it challenging to scale up. Data-driven AI methods have demonstrated strong adaptability and robustness across various scenarios in multiple fields, making them a promising solution to address the fragility of UWB positioning. Most existing UWB positioning algorithms rely on AI assistance with model-specific designs. This paper proposes a unified end-to-end Long Short-Term Memory (LSTM) network model, which achieves universal adaptability to three UWB positioning models (TOA, TDOA, and AOA) and thus proves the adaptability of pure AI models to different UWB positioning approaches.Comprehensive simulations and real-world experiments show that the LSTM-based end-to-end model significantly outperforms physics-based calculation methods under non-ideal conditions such as coplanar anchor deployment. Moreover, it enables a unified model to solve different positioning models, highlighting the broad potential of data-driven learning approaches in indoor positioning."