WiFi has become the dominant access method for the mobile Internet. However, its performance, safety, and privacy are far from satisfying, and its potential utility for studying human behavior is far from fully explored. WiFi Union project aims to achieve better WiFi performance, safety, privacy, & utility. Intelligence is the key to achieve these goals, and our approach is to obtain the intelligence needed through sharing and machine learning in a union of various parties in the WiFi world, including users, residential AP owners, SOHO AP owners, enterprise WiFi operators, commercial WiFi operators, AP manufacturers, chip manufacturers, smart phone manufacturers, mobile operating system manufacturers.
Towards above goal, we start our long-term WiFi Union project which explores various directions to how coordinated and unionized efforts among different WiFi parties can help us go step by step towards a better WiFi world, a union of all WiFi parties.
Our WiFi data collection includes WiFi data from a university campus with about 60K people, 3000 APs (and growing), and a couple of popular university Apps with more than 10K downloads, an WiFi AP testbeds with 100+ (and growing) deployed OpenWrt-based APs in the wild, a testbed with 300 OpenWrt-based APs, a couple of dozens of Android smart phones, a couple of dozens of Raspberry Pis, and a kit (one cisco AC, 7 cisco APs) for providing WiFi-based Internet access for large events (e.g. those held in Tsinghua Main Auditorium)
- Measurement, diagnosis and optimization for enterprise networks such as university campus networks
- Measurement and Optimization for autonomous WiFi networks: Measuring and optimizing packet-level latency, interference across autonomous WiFi networks, e.g. residential WiFi networks, SOHO WiFi networks
- Performance measurements and Intelligent AP selection for public WiFi APs.
- SAVY: Scoring WiFi AP’s vulnerabilities. We build a mobile AP that answer the following question: “How vulnerable is the Public WiFi AP you are using”?
- Social events as observed via WiFi measurements
- Performance analysis for very dense WiFi networks
- Understanding and Improving trajectory privacy in WiFi data
- The correlation of user behavior (as observed via WiFi) with academic/research performance