Signal strength-based localization is becoming increasingly important as a complementary approach to GNSS which can be disrupted or is entirely unavailable indoors. This session presents advances across Bluetooth/WLAN RSSI, UWB, CSI, and magnetic field-based positioning methods. The papers employ modern machine learning and signal processing techniques, including Bayesian and graph neural networks, tensor decompositions, and compressed sensing, to approach various challenges: multipath mitigation, hardware impairments, correlated outliers, uncertainty quantification, and efficient map representation. Diverse applications ranging from indoor zone localization and asset tracking to railroad odometry show that with the proposed techniques, reliable navigation can be achieved across diverse environments.
