9. Tracking

Keeping track of motion in the physical world is a crucial part of any VR system. Tracking was one of the largest obstacles to bringing VR headsets into consumer electronics, and it will remain a major challenge due to our desire to expand and improve VR experiences. Highly accurate tracking methods have been mostly enabled by commodity hardware components, such as inertial measurement units (IMUs) and cameras, that have plummeted in size and cost due to the smartphone industry.

Three categories of tracking may appear in VR systems, based on what is being tracked:

  1. The user's sense organs: Recall from Section 2.1 that sense organs, such as eyes and ears, have DOFs that are controlled by the body. If a display is attached to a sense organ, and it should be perceived as in VR as being attached to the surrounding world, then the position and orientation of the organ needs to be tracked. The inverse of the tracked transformation is applied to the stimulus to correctly ``undo'' these DOFs. Most of the focus is on head tracking, which is sufficient for visual and aural components of VR; however, the visual system may further require eye tracking if the rendering and display technology requires compensating for the eye movements discussed in Section 5.3.
  2. The user's other body parts: If the user would like to see a compelling representation of his body in the virtual world, then its motion should be tracked so that it can be reproduced in the matched zone. Perhaps facial expressions or hand gestures are needed for interaction. Although perfect matching is ideal for tracking sense organs, it is not required for tracking other body parts. Small movements in the real world could convert into larger virtual world motions so that the user exerts less energy. In the limiting case, the user could simply press a button to change the body configuration. For example, she might grasp an object in her virtual hand by a single click.
  3. The rest of the environment: In the real world that surrounds the user, physical objects may be tracked. For objects that exist in the physical world but not the virtual world, the system might alert the user to their presence for safety reasons. Imagine that the user is about to hit a wall, or trip over a toddler. In some VR applications, the tracked physical objects may be matched in VR so that the user receives touch feedback while interacting with them. In other applications, such as telepresence, a large part of the physical world could be ``brought into'' the virtual world through live capture.

Section 9.1 covers the easy case of tracking rotations around a single axis to prepare for Section 9.2, which extends the framework to tracking the $ 3$-DOF orientation of a 3D rigid body. This relies mainly on the angular velocity readings of an IMU. The most common use is to track the head that wears a VR headset, but it may apply to tracking handheld controllers or other devices. Section 9.3 addresses the tracking of position and orientation together, which in most systems requires line-of-sight visibility between a fixed part of the physical world and the object being tracked. Section 9.4 discusses the case of tracking multiple bodies that are attached together by joints. Finally, Section 9.5 covers the case of using sensors to build a representation of the physical world so that it can be brought into the virtual world.

Steven M LaValle 2016-12-31