AR Based Indoor Navigation

Left: The fused output of the navigation system visualized in a VR fashion and simulating the user trajectory in the building digital twin.

Right: The actual video feed when this trajectory was recorded.

Trajectory length: 70+ meters including the stairs.

Note: The shown markers in this video are not used for localization! they are used as anchor points only for evaluating the navigation system.

My master's thesis project was the result of a cooperation between the NavVis Navigation Team and TUM. The research is investigating the viability and practicality of using visual techniques incorporated with 3D building structure knowledge for indoor navigation. The proposed system is a hybrid of absolute visual positioning and relative visual tracking. The positioning is done using image based poses estimation approach relying on dense feature matching (using deep neural network) and the building's 3D structure. While the tracking is done with the help of Goolgle ARCore. This video shows the result of the strategic fusion between the absolute and relative data.

All the indoor scans were acquired using the NavVis M6 Trolley, one of the state of the art mobile scanning devices in the industry.

Augmented Reality App using Nokia HAIP Technology

Hi everyone one,

Today I present you my latest achievement in the field of Augmented Reality.

In the last post I have presented a way using libGDX to augment a box in the right orientation.
This solution is location invariant. Which means whenever I change my location the box will change its location with me. There is no sense of displacement.

Today, using the Nokia HAIP (High Accuracy Indoor Positioning) technology, I was able to add the sense of position to the augmented objects.

Here is a video showing the result we got  and show the used devices and technologies

Note: HAIP is based on BLE Bluetooth Low Energy to get the location of objects/people.