WIFI-based Crowd Safety System

The widespread adoption of mobile devices allows services based on an indoor positioning system to be implemented [1]. Understanding your customers and their behaviors is a key point of a company. The analysis of wireless signals emitted by portable devices such as smartphones, laptops, and tablets enables the extraction of positional data from those devices. In public locations, security measures such as emergency routes can be enhanced by analyzing crowd behavior.

The idea of this thesis it to provide an implementation of a real-time crowd monitoring system to ensure physical distancing in crowds using WIFI sensors. The advantages of using WiFi sensors is that it offers a significantly more cost-effective for wide area coverage than comparable systems (for example, a camera-based crowd monitoring solution). Also, it is significantly easier to install and configure than comparable systems (for example, a camera-based crowd monitoring solution).

Henceforth, this thesis involves answering challenges (which can be adjusted depending on the type of thesis - BA/MA/MP) such as:

• Is it possible to create a real-time alerting system using WIFI sensors that is sufficiently close to real-time to be effective for crowd safety management?
• What is the optimal density of sensor coverage required for such a system? In other words, what is the optimal number of WIFI sensors per 100 square meters?
• Are there any constraints or requirements for the installation of the WIFI sensors? For example, what is the optimal layout of the WIFI sensors at the venue?
• What is the best algorithm for calculating the density thresholds in terms of signal strength and the number of signals captured?

Tasks on an engineering aspect demands:

• Implement a streaming data pipeline from the sensor fleet to the application.
• Filter out WIFI signals that are not relevant. For example, beacon frames, multicast MAC addresses, and OUIs that do not match smartphone manufacturers.
• Visualize in near real-time the crowd density values for each sensor location on the floor plan of the venue.
• Implement an alerting mechanism for sensors whose crowd density value exceeds a predefined safety threshold.
• The relevant application parameters should be configurable. E.g., the position of each sensor, crowd density thresholds, etc.