Learning Displacement-Aware WiFi Representations for Weakly Supervised Relative Localization

In three linesRelative WiFi localization without dense coordinate annotations. Intersection Pathway aligns WiFi fingerprint traces and inertial motion vectors in a shared additive latent space, enabling direct relative-displacement inference. Validated on synthesized dataset from real measurements.

Summary generated by Claude — human-verified