Google Research published a study showing that coordinated routing through navigation apps can improve traffic speeds and lower emissions in major U.S. cities.

The study builds on earlier work such as Project Green Light, which used artificial intelligence to optimize traffic signal timing. While navigation services already provide individual route suggestions, system‑wide coordination has not been widely implemented.

In a six‑month experiment, the Google Maps algorithm was modified to favor alternative routes with similar travel times for trips that entered pre‑selected congested road segments. The intervention was applied to all trips in ten cities, affecting less than 2 % of observed trips.

Cities were selected based on congestion levels and availability of ground‑truth data. Approximately 100 road segments with recurring bottlenecks were identified for the study.

Results were analyzed using a hierarchical Bayesian model that considered both city‑wide and hourly effects. Across the treated segments, average driving speeds increased by about 2 %. On all segments impacted by the rerouting, speeds rose by roughly 0.35 % to 0.5 % during peak morning and afternoon periods. Fuel consumption decreased by 0.5 % to 1.0 % on the targeted segments. The estimated CO2 reduction reached thousands of tons per city annually.

The findings indicate that diverting a small fraction of trips away from major bottlenecks can raise speeds and lower emissions across the broader network. Both navigation users and non‑users benefit from reduced travel time and lower emissions.

The research demonstrates a method for conducting large‑scale, experiment‑based traffic management using existing navigation platforms. It provides a framework for future work on dynamic signal control and real‑time network optimization as smart‑city infrastructure develops.