MAC randomization mitigation refers to the set of techniques and strategies used to counteract the effect of MAC address randomization, a privacy feature in modern mobile devices that causes them to broadcast randomly generated MAC addresses rather than their true hardware identifiers, making it difficult to uniquely track or count individual devices over time. This challenge matters significantly for WiFi-based sensing research because systems that rely on probe request sniffing for crowd counting, occupancy estimation, or movement tracking can produce severely inflated or inconsistent device counts when the same physical device appears under multiple MAC addresses. Key mitigation variants include fingerprinting devices using non-MAC probe request attributes such as information elements, signal strength patterns, and sequence numbers, as well as applying deep learning classifiers or clustering algorithms to group probe requests that likely originate from the same device despite changing addresses.
Source Papers
- A low-cost automatic people-counting system at bus stops using Wi-Fi probe requests and deep learning ↗ — A low-cost automatic people-counting system at bus stops usi
- Estimating indoor crowd density and movement behavior using WiFi sensing ↗ — Estimating indoor crowd density and movement behavior using