MIT has developed a GPS system that uses AI to recognize roads based on satellite photos.
GPS systems have been accompanying us for many years, we are used to its use, and its pros and cons. The latter seem to not matter much to us, and we live with them without problems, such as the lack of precision of the system when we talk about precise locations. Now the Massachusetts Institute of Technology , which we usually know from MIT, has developed, together with researchers from the Institute of Computer Research of Qatar, they have developed a GPS system with Artificial Intelligence Qatar a revolutionary GPS positioning system that is capable of increasing accuracy to unknown limits now. And all thanks to the AI, which will be able to better identify our location, with more precision.
It will be able to detect even the lane through which we circulate
Surely you will have noticed sometime, that GPS is not able to locate us well within the road. If this usually gives us the indications, it is because he knows that we are going in a certain direction, and that therefore we will have to take a certain exit in one direction, but because he detects exactly the place where we are. This new technology is able to improve that positioning and make it much more precise from artificial intelligence. The new software they have created is called RoadTagger , and uses a mixture of maps and artificial intelligence to better identify the peculiarities of each of these places.
What is achieved with this software, for example, is to recognize exactly how many lanes there are at the point we are traveling through, this logically favors more precise indications. Since the neural networks of this software can know thanks to the analysis of satellite photographs of that location the special conditions of the road. And so with many other aspects of the maps, which would otherwise be very difficult to see. The accuracy of this system is quite high, since in the tests carried out up to 77% of the lanes of the different roads were correctly recognized . In addition, the number of roads recognized correctly rose to 93%.Therefore, it is a system that based on satellite images that can recognize the peculiarities of roads in a more precise way than the current GPS, because it analyzes the images of the places through which we travel, so it is based on information more tangible
The maps created are the result of the union of the OpenStreetMap and Google Maps open system . According to its creators, this type of GPS navigation system is much more practical than those now used by large companies such as Google Maps, which base this type of terrain details on the sweeps of their cars that record the highways and roads on which we travel In this way, only satellite images are enough to have the necessary information on the tracks. Even in the future it could serve to tell us if there is parking available near us, thanks for example to the analysis of images of parking areas, in turn analyzed through the neural networks of artificial intelligence.