One of the consistent challenges to Google Maps now a day is figuring out the right direction to go using the blue dot. Many times app instructs us to go south direction, but mind wonders in the north. Why does that happen? Are we not sure what is App is guiding? Where did trust fly?
Over the years Google is attempting to improve the accuracy with tools like Compass and GPS but, those are having their physical limitations makes challenges tough especially in urban environments.
Maps with ARCore
Last year, Google announced plans to come up with augmented reality(AR) to Google Maps. On February 11, at least for a couple of users that became real. Google has rolled out AR Maps to a small subset of Google Local Guides, people who voluntarily add images, reviews, and helps to improve the App.
The feature is designed for pedestrians to assist them to navigate around coroners of cities by showing them digital street signs and virtual arrow signs covered on the footway to indicate which way to walk.
To check this in real, Users can hold their phones up in front of them, and the AR version of Map is activated. It assumed to solve the problem many people run into when using a Map to navigate: which turn to take, less traffic. To reduce accidents, screen popups encourages users to abandon phone use after shown the correct path and focus more on moving.
VPS, Street View
Visual Positioning Service plots the location of a device based on representation rather than GPS signals. VPS first generates a map by taking a series of images which have known the area and analyzing them for key-visual characteristics, such as an outline of objects, create large scale and fast searchable index of those optical characteristics. To localize the device, VPS compares the features in the imagery from the phone to the VPS index, and this is where Street View comes in picture.
To help people explore the world more quickly, deeply, over ten years ago, Google launched Street View. In the time, Street View has continued to expand its coverage of the world, empowering people not only preview their route but also step inside famous landmarks and museums, restaurants. To deliver global localization with VPS, Google connects it with Street View data which will help more accurately determine the position of a device and guide people towards the target.
Although this approach works well in theory, making all well in practice is challenging to Google. Locations and imagery from the phone may differ from what the scene looked like when Street View imagery get collected. Explaining this scenario Tilman Reinhardt, Software Engineer, Google Maps says;
To overcome this problem like dynamic light movement, constructions that are likely transient Google claims that this is just one of many ways they use machine learning and AI to improve accuracy in Maps.