Graphical abstract of the system architecture
Dashboard in the Grafana visualization platform
The sensors periodically upload crowding measurements with the number of devices detected to a cloud server, which ingests them and then renders the collected information on a data visualization platform. Users can then access this platform to visualize the information collected by the sensors on customizable dashboards with geo-temporal renderings of the crowding information.
Sensors can use multiple uplink options that mitigate network limitations at the sensor installation location. The sensors periodically upload crowding measurements with the number of devices detected to a cloud server, which ingests them and then renders the collected information on a data visualization platform. Users can then access this platform to visualize the information collected by the sensors on customizable dashboards with geo-temporal renderings of the crowding information.
Sensors can use multiple uplink options that mitigate network limitations at the sensor installation location. If Wi-Fi coverage is not available, LoRaWAN (Long Range Wide Area Network) can be used instead.
This crowd detection architecture is GDPR (General Data Protection Regulation) compliant as no private information is stored, i.e. sensors only send the number of detected devices to the server at regular intervals.
Raspberry Pi -> Coordination and Processing
Wi-Fi dongle -> detection of Wi-Fi devices
Bluetooth dongle -> detection of Bluetooth devices
LoRa board for communication via LoRaWAN protocol (if required)
Wi-Fi enabled sensor version (without refrigeration)
LoRa enabled sensor version (with refrigeration)
Mounting a crowding sensor