Our initial implementation focused on building a scalable backend architecture and edge interface on top of a proprietary edge data generator to ensure that the rest of our platform could bear the load of hundreds of thousands of sensors. As a result our primary engineering focus has been on backend functionality rather than UI. Here are the areas we have spent the most time implementing:
- Dynamic, truly random pedestrian and vehicle data generator. Capable of generating data at >50sensors/second. Tested on both simulated edge devices (rPi3) and Cloud Servers (AWS EC2)
- Edge to Cloud data pipeline leveraging MQTT. Capable of 20+ sensors/second in a single thread.
- Backend Analytics Engine. Conversion from raw data stream into transformable data. Handles analysis and calculation of metrics such as congestion.
- End-to-End Application handling, processing, and visualizing real-time data from 860 simulated sensors.