DarLoc: Deep learning and data-feature augmentation based robust magnetic indoor localization

Qinghu Wang, Jie Jia, Yansha Deng, Jian Chen, Xingwei Wang, Min Huang, Abdol-Hamid Aghvami

Abstract: Highlights•We investigate a magnetic localization system with multiple heterogeneous factors.•We propose a feature extraction method to handle the challenges of various speeds.•We propose data and feature augmentation methods to model various speed patterns.•We propose a multi-scale spatial–temporal learning model for accurate localization.