From aggregated and anonymized location data to predictions based on historical data and machine learning: this is how Google Maps calculates traffic and travel times by car. And with the integration of Gemini the interaction is increasingly natural
Google Maps or Waze. These are the two navigators used today by the majority of motorists, favored by the now widespread presence in modern cars of screens and systems similar to those of smartphones (Android Auto and Apple CarPlay). Waze, also owned by Google, is growing but remains less well-known, while it is more difficult to find someone who has never consulted Google Maps. The American software seems magical at first glance: it identifies the quickest and most important route “reads” the trafficcoloring the lines from green to red depending on the level of congestion. How does he do it?
TRAFFIC AND TIMES
—
The Mountain View giant itself has taken care of explaining the mechanisms of Maps for reading traffic in real time. The main “secret” it is the great diffusion of the application: being able to count on a huge portion of users who use the service at that moment, Google, through the internet connection of smartphones, manages to aggregate a lot of anonymous data. These have a double utility. On the one hand they are used for determine congestion on the roads in real timein order to propose possible alternative routes to other motorists in transit. On the other hand, however, they are archived so that they constitute a “historical” basis for that stretch on that day and at that time. Crossing the two elements with advanced techniques machine learningthe application can estimate the conditions that the motorist will encounter upon arriving at a certain stretch, combining current conditions and historical trends. This results in an accurate prediction of travel times, used to calculate routes and arrival times. Information from third parties, such as local authorities or national departments, also contributes to the accuracy of Google Maps, without forgetting the reports from motorists which, with a few taps on the car screen, can signal the presence of accidents or maintenance interventions on the road. In this way, travel times are in most cases very reliable, given that the effectiveness of the system depends on the amount of data available and the network connection: in areas with little traffic or in case of sudden events, the estimate may take a few minutes to update.
GEMINI
—
Traffic is a well-established competitive advantage of Google Maps. In some time it is likely to become so too GeminiGoogle’s virtual assistant with artificial intelligence which, update after update, takes on an increasingly central role in the experience of using Maps. Gemini in the car it is a recent introduction, and Google is progressively expanding its functions. Voice control of navigation was already possible with the predecessor Google Assistant, but with the integration of Gemini the interaction has become more natural and conversational, allowing more complex and contextual requests. Gemini can search through details of places along the route and reviews to answer complex questions, such as: “Hey Google, I’m craving pizza. Are there any good pizzerias open near my destination?” The dialogue with the intelligent assistantFurthermore, it can continue without interruptions, so if one of the proposals seems promising you can investigate further by asking, for example, if the place accepts animals or what the customers who have left a review think about it. The goal is to make the interaction more fluid and spontaneous, reducing the need to touch the screen so as to leave the driver to the most important task: driving.
© ALL RIGHTS RESERVED
