Fusion of Wi-Fi and Light Data for Detecting Companion-Based Shopping Behaviors in Indoor Retail Environments

The growing need for personalized and privacy-preserving indoor services in shopping malls has driven the development of accurate and intelligent localization techniques. This paper introduces a novel system that fuses Wi-Fi signal fingerprinting with light sensor data, namely AP-Light, to infer user proximity and group relationships, offering precise localization without relying on privacy-invasive camera inputs. By leveraging light sensor data, the system effectively identifies social contexts, such as detecting companion-based shopping behaviors. Furthermore, the integration of this location and relationship information with large language models (LLMs) enables the dynamic generation of personalized, context-aware advertising. For example, promotions like "Buy 1 Get 1 Free"can be tailored for users shopping with companions. This fusion-based methodology not only enhances localization accuracy and preserves user privacy but also transforms retail environments into intelligent, context-aware spaces, delivering real-time, tailored engagement. Experimental evaluations demonstrate the system's robustness and potential to redefine shopping experiences in indoor retail settings.