Wi‑Fi/GSM Geolocation in Large Cities Using ML for 30 Million Users
summary
Uses machine learning to provide Wi‑Fi/GSM geolocation for 30 million users in large cities by building and using millions of 2D signal-strength maps of Wi‑Fi access points and cell towers. Combines statistical methods to compute accurate signal-strength maps with engineering techniques to generate and serve hundreds of millions of such maps at scale.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 2:12GPS positioning problems
- 4:01Wi-Fi signal idea
- 9:26Signal propagation modeling
- 12:34Bayesian model errors
- 15:23Wi-Fi model building
- 17:30Signal ensembling
- 18:13Cell tower usage
- 18:55Outliers and clusters problems
- 21:05Data updates
- 23:13Results and metrics comparison
- 26:41Engineering implementation
- 28:10Q&A
related talks
06mMIMU: Mobile WiFi Usage Inference by Mining Diverse User Behaviors
07mMIFF: Human Mobility Extractions with Cellular Signaling Data under Spatio-temporal Uncertainty
30mDisambiguating Explicit and Implicit Geographic References...
06mMachine Learning for Phone-Based Relationship Estimation: The Need to Consider Population ...
19mHow we leverage AI to build great maps at massive scale
07mHERMAS: A Human Mobility Embedding Framework with Large-scale Cellular Signaling Data
14mLeveraging ML to Detect Application HotSpots [@scale, of Course!]
06mCellPred: A Behavior-aware Scheme for Cellular Data Usage Prediction