Graph-Based Floor Separation Using Node Embeddings and Clustering of WiFi Trajectories

Name: Graph-Based Floor Separation Using Node Embeddings and Clustering of WiFi Trajectories
Uploaded: 2026-10-06T22:57:09.769228+00:00
Duration: 320 s
Description: The research developed a data-driven framework to automatically separate floors in multi-story buildings using Wi-Fi signal strength data. By treating user movements as a graph, it used Node2Vec embeddings to capture the building's vertical structure and K-Means clustering to assign trajectories to specific floors without needing prior building maps.

The research developed a data-driven framework to automatically separate floors in multi-story buildings using Wi-Fi signal strength data. By treating user movements as a graph, it used Node2Vec embeddings to capture the building's vertical structure and K-Means clustering to assign trajectories to specific floors without needing prior building maps.

Open the full episode