Combinatorial Data Augmentation for Real-Time Indoor Positioning: Concepts and Experiments

Precise positioning has become one core topic in wireless communications by facilitating candidate techniques of beyond 5G and 6G. Nevertheless, most existing positioning algorithms, categorized into geometry-driven and data-driven approaches, fail to simultaneously fulfill diversified requirements for practical use, e.g., accuracy, real-time operation, scalability, maintenance, etc. This article aims at introducing a new principle, called combinatorial data augmentation (CDA), a catalyst for tightly integrating geometry and data-driven approaches. We first explain the concept of CDA and its critical advantages over the two standalone approaches, followed by validating its effectiveness by field experiments with WiFi round-trip time and inertial measurement units.