In addition to academic papers such as [83], some of the most useful coverage for IMU calibration appears in corporate white papers, such as [248]. For magnetometer calibration, see [93,157,167,336]. Oculus Rift 3D orientation tracking is covered in [168,166,167,170]. To fully understand vision-based tracking methods, see vision books [112,193,323]. Many approaches to PnP apear in research literature, such as [362,375]. An excellent, but older, survey of tracking methods for VR/AR is [352]. One of the most highly cited works is [143]. See [239] for integration of IMU and visual data for tracking.
Eye tracking is surveyed in [64,346]. Human body tracking is covered in [376]. To fully understand kinematic constraints and solutions to inverse kiematics problems, see [8,10,48]. SLAM from a robotics perspective is thoroughly presented in [328]. A recent survey of SLAM based on computer vision appears in [87]. Filtering or sensor fusion in the larger context can be characterized in terms of information spaces (see Chapter 11 of [165]).
Steven M LaValle 2020-11-11