Packet Allocation: Improving the Accuracy of Mobile Device Count Estimation

This study specifically examines development of a method for estimating the number of mobile devices in dense smoke environments that arise because of tunnel fire accidents, with the aim of facilitating evacuation. Various sensing technologies such as radio wave positioning and image processing have been used with inflow analysis systems, but their applicability in environments with poor visibility and confined spaces, such as smoke-filled tunnels, remains limited. To address this shortcoming, we propose a method using Wi-Fi probe requests for estimation of the number of mobile devices. Prior research has demonstrated that dense smoke exerts only minimal effects of on radio wave characteristics in such environments. Building on these earlier findings, we apply principal component analysis to reduce the dimensionality of features extracted from Wi-Fi probe requests and subsequently apply cluster analysis to estimate the number of devices. To enhance the estimation accuracy of our proposed method, we introduce a novel packet data allocation technique based on data collected from multiple Wi-Fi packet captures deployed within a tunnel environment. Then we validate the accuracy of this allocation technique and report the effectiveness of our proposed method for estimating the number of mobile devices.