Abstract: Deep learning methods have been widely applied in indoor localization to mitigate the impact of non-line-of-sight (NLOS) and multipath effects, achieving high-precision results. However, ...