Virtual Reality has become one of the biggest (if not the actual biggest) trends in technology, mobile app development, and training lately: with over one million units sold in 2017, (combining Playstation VR, Oculus Rift, and HTC Vive) Virtual Reality started to gather a major following in the gaming market (as expected) but not just there.
Since Virtual Reality and Augmented Reality became “market ready”, this technology wanted to conquer other markets: for example, VR headsets are currently used to train surgeons, pilots and many more.
Today, VR is reaching the intersection with another big technology which saw its power increasing drastically during the last year: Machine Learning.
Let’s try and break down the process in a simpler way:
What Is SLAM?
Simultaneous Localization And Mapping is “is the computational problem of constructing or updating a map of an unknown environment while simultaneously keeping track of an agent’s location within it.” (from Wikipedia).
How does this technology apply to Virtual Reality then? Well, basically SLAM is triggered by the VR’s peripheral, who keeps track of the surrounding area while translating these pieces of information in the virtual environment and vice-versa.
What happens, indeed, is a machine learning-based process that automatically elaborates all the projections our headset is gathering from the surrounding environment into a solid, precise representation that is translated in the actual virtual representation.
Of course, current headsets don’t have the computing power to elaborate all these data by themselves, so, for now, all these data are elaborated from another machine (that is required in any case to run the peripheral)
To reference probably the most important one, Microsoft’s HoloLens is currently the representation of how SLAM could be implemented in an Augmented Reality-based environment: not only because they are able to connect digital content (such as graph projections and so on) to the real, physical environment, but also because of the amount of potential this technology has currently.
Imagining the near future, HoloLens has a lot of possible usages, from engineering to development, passing through proper office-based tasks.
On another level, mapping in real time without relying on pre-existing pieces of information (like Google Maps, for example), could be great if applied to cars’ infotainments:
Tesla is already relying on SLAM in its basic form and application since their autonomous driving is structured so it can automatically compile the sensors’ data that is constantly split testing against their map’s data.