My whole purpose is for indoor positioning.
For example: I don't need the SLAM modeling part of it. My understanding is that the infared camera is used for 3-D mapping of the environment. Does it feed into the VIO algorithm? Can the VIO work with the regular camera?
Project Tango's motion tracking technology uses only IMU and a grayscale wide angle camera to do the VIO. Depth camera is an additional sensor which provides point clouds and has nothing to do with Motion Tracking.