BaG: Behavior-Aware Group Detection in Crowded Urban Spaces Using WiFi Probes

Group detection is gaining popularity as it enables variousXzX applications ranging from marketing to urban planning. Existing methods use received signal strength indicator (RSSI) detect co-located people groups. However, this approach might have difficulties in crowded spaces since many strangers with similar mobility patterns could be identified Moreover, RSSI vulnerable factors like the human body attenuation and thus unreliable scenarios. In work, we propose a behavior-aware group system (BaG). BaG fuses people’s information smartphone usage behaviors. We observe that tend phone patterns. Those effectively captured by proposed feature: number of bursts (NoB). Unlike RSSI, NoB more resilient environmental changes only cares about receiving packets or not. Besides, both correspond same underlying grouping information. method based on collective matrix factorization reveal hidden associations factorizing simultaneously. Experimental results indicate outperforms baseline approaches $3.97\% \sim 15.79\%$ F-score. The also achieve robust reliable performance scenarios different levels crowdedness.