Identifying Stops and Moves in WiFi Tracking Data.

Area of Science:

• Computer Science
• Data Science
• Ubiquitous Computing

Background:

• Classical individual tracking methods (GPS, video surveillance) are costly and do not scale for large populations or urban areas.
• WiFi tracking presents a scalable, cost-effective, and unobtrusive alternative for large-scale crowd monitoring.
• WiFi data presents challenges due to low positional accuracy and sparse, irregular signals, complicating trajectory analysis.

Purpose of the Study:

• To evaluate the applicability of existing GPS-based stop/move detection algorithms to WiFi tracking data.
• To identify and propose improvements for WiFi tracking data analysis, specifically for segmenting trajectories into stops and movements.
• To facilitate deeper insights into crowd behavior and visited locations using WiFi tracking data.

Main Methods:

• Investigated three established algorithms for segmenting GPS trajectories into periods of stops and movements.
• Assessed the performance and adaptability of these algorithms when applied to WiFi tracking datasets.
• Analyzed the unique characteristics of WiFi tracking data to inform algorithm refinement.

Main Results:

• Demonstrated that GPS-based algorithms can be adapted for WiFi data, though performance varies.
• Identified specific challenges in applying GPS algorithms directly to lower-accuracy WiFi data.
• Established a baseline for stop/move detection in WiFi trajectories, highlighting areas for enhancement.

Conclusions:

• The segmentation of trajectories into stops and movements is crucial for extracting meaningful information from WiFi tracking data.
• Existing GPS-based methods provide a foundation but require adaptation for optimal WiFi data analysis.
• Further research and algorithm improvements are necessary to fully leverage WiFi tracking for understanding large-scale human mobility and behavior.