Google Maps published a a blogpost on Thursday on traffic and routing to explain to people how it identifies a massive traffic jam or determines the best route for a trip.
While Maps can easily identify traffic conditions using the aggregate location data, the data still is not sufficient to predict what traffic will look like 10, 20, or 50 minutes into a user's journey. According to Google Maps, that's where technology really comes into play. To make such predictions, Maps analyzes historical traffic patterns for roads over time. The combined database of historical traffic patterns with live traffic conditions using machine learning helps generate predictions.
Google Maps revealed that it recently partnered with DeepMind, an Alphabet AI research lab, to improve the accuracy of its traffic prediction capabilities. Maps also states that its predictions have been ''consistently accurate for over 97% of trips.''. By working with DeepMind, Maps has been able to cut the inaccuracy percentage even more and for that, it has used Graph Neural Networks, a machine learning architecture. The technology has helped make significant improvements in places like Berlin, Jakarta, São Paulo, Sydney, Tokyo, and Washington D.C.
Moreover, with the recent significant changes in traffic patterns around the world that the Coronavirus pandemic has brought, Google Maps has had to update its models to become more agile.
Google Maps' predicted traffic models are also a major part of its driving routes predictions. Maps automatically finds users lower traffic alternatives as soon as it predicts that traffic is likely to become heavy in one direction. In addition to that, many other factors also play their role, such as road quality. If Maps identifies a road as difficult to drive down, it is less likely to recommend it to a user as part of their route. Similarly, Maps also looks at the size and directness of a road, for instance taking a highway is likely to be more efficient than taking a smaller route with several stops.
Maps has two other important sources of information to help recommend users the best route to take: authoritative data from local governments and real-time feedback from users. Authoritative data helps Google Maps learn about speed limits, tolls, or whether a road is restricted. Feedback from users is helpful too. For instance, an incident report from drivers would help Maps know that a lane or road is closed, or whether a road is restricted because of construction.
In conclusion, with Google Maps' traffic predictions combined with live traffic conditions, it can automatically reroute a user, hence helping them reach their destination conveniently and on time.
Maps states that predicting traffic and determining routes is a complex, and that it will keep working on tools and technology to make users' travelling experience as safe and efficient as possible.