听
Smartphones and route-planning apps have changed the way we navigate, guiding us from A to B with maximum efficiency. But what if we want to take the scenic route?
Soon they may help us do that, too. An AI system called Autobahn can use Google Street View images to optimise your route for a particular type of scenery, such as mountains or views of the ocean.
Advertisement
, Germany, and and colleagues at Belgium鈥檚 Hasselt University first divided the landscape into squares with 1-kilometre sides and downloaded the Google Street View panorama from a spot on the largest road within each square.
Each square was assumed to equate to around 1 minute鈥檚 driving time, assuming an average speed of 96 kilometres per hour. Next, a type of AI called a neural network classified the Google Street View鈥檚 scenery as forest, mountain, fields, water, a sight (like a monument) or non-scenic.
To plan a route, you simply provide a start and end point, a maximum travel time and a scenery preference. If you choose mountains, for example, the AI chooses the route with the most mountain views.
No blocked views
Previous approaches to scenic route planning have used maps instead of Google Street View. So they can tell you how to drive geographically closest to the ocean, but when you get there, you might find a wall or tree blocking your view.
The researchers tried Autobahn for the Spanish island of Majorca; Rhone-Alpes in France; and Santa Barbara in California 鈥� 鈥渢he preferred holiday areas of the authors鈥�, says Sch枚ning.
To test it, the team carried out an online survey using images from journeys in these three settings. The scenic Autobahn-generated route was compared to the fastest route and 24 volunteers were asked to say which route they would prefer to take. Most preferred the Autobahn route.
The pair will present the research in Sonoma, California, in March.
at the University of Florida says the system is fast, effective, applicable to cell phones and easily scalable to large areas. 鈥淚t鈥檚 a stepping stone towards a more comprehensive route planning system that would know automatically what you want without having to specify any parameters 鈥� something that can read your mind based on your past history or the weather,鈥� Hochmair says.
鈥淚n five, 10, 15 years, when we all drive autonomous cars, I don鈥檛 believe we鈥檒l still go the shortest paths from A to B,鈥� says Sch枚ning. With driving outsourced to machines, maybe we will be free to enjoy the scenery, he adds.
Journal reference: ACM IUI 2016, DOI: 10.1145/2856767.2856804
