
Michael Kopp (center) with Giovanni Lanfranchi (right), chief technology officer at HERE Technologies, and Ina Fried (left), chief technology correspondent at Axios.
A recent article by ITS International showcases the power of Artificial Intelligence (AI) in enhancing mobility by predicting traffic and managing vehicle fleets. The article summarizes “Location-Based AI Enabling a Mobile and Efficient Future” session at CES 2020 in Las Vegas. Michael Kopp, a founding director of IARAI and director of data science at HERE Technologies, discussed in this session how training AI on geospatial data can improve traffic in cities and keep location data private.
Michael Kopp showed that location-based AI can effectively predict traffic in the Traffic4cast competition. Traffic4cast 2019 was based on industrial dataset with over 100 billion points provided by HERE Technologies. Traffic data was collected in Berlin, Moscow, and Istanbul by a large fleet of probe vehicles. The data was sampled on 100m x 100m squares every 5 minutes throughout a year. The high resolution traffic trajectories were mapped into spatiotemporal cells and summarized as movies; traffic volume, speed, and direction were encoded as red, green, and blue color channels. The competition results demonstrated that AI can provide insights into complex traffic systems and offer solutions for improving traffic.