Spatio-Temporal Feature Learning-Based WiFi Fingerprinting Localization

WiFi fingerprinting localization has become one of the most promising indoor localization techniques over the past two decades, due to its cost-effectiveness and widespread deployment. Traditional methods, however, are susceptible to inherent wireless signal noise, leading to degraded accuracy and poor robustness in localization. This chapter provides a comprehensive overview of state-of-the-art approaches that leverage spatio-temporal features in sequential fingerprint data to alleviate this problem. It systematically elaborates on three main categories of existing methods, i.e., the traditional trajectory optimization method, spatio-temporal feature learning with single-modal data, and spatio-temporal feature learning with multimodal data, all of which are representative solutions for enhancing the performance of WiFi fingerprinting localization systems.