The »Roadlytics« project deals with traffic congestion caused by logistics vehicles in Hamburg’s city center. The goal is to collect geo-referenced data for loading and delivery traffic using so-called IoT sensors. This data is used to generate learnings on how to avoid those congestions.
To do this, the team led by product owner Elisa Soncin used artificial intelligence and machine learning techniques to identify stop hotspots, analyzed their influence on traffic and planned new routes and stress-free parking for logistics vehicles. A special model, the unsupervised learning model — Density-Based Spatial Clustering with Noise (DBSCAN), is used for clustering geo-referenced data to predict stopping hotspots with great accuracy.