Adaptive Complementary Filter for Hybrid Inside-Out Outside-In HMD Tracking With Smooth Transitions

Area of Science:

• Computer Science
• Human-Computer Interaction
• Robotics

Background:

• Head-mounted displays (HMDs) in virtual reality (VR) commonly use inside-out tracking (visual SLAM).
• Outside-in systems (e.g., motion capture) offer alternative tracking but are prone to occlusion and blind spots.
• Hybrid tracking combines inside-out and outside-in methods to overcome individual limitations.

Purpose of the Study:

• To develop a sensor fusion algorithm for hybrid HMD tracking that ensures smooth transitions between tracking modes.
• To mitigate user discomfort caused by tracking discontinuities when outside-in signals are recovered.
• To provide a configurable solution balancing user experience and tracking accuracy.

Main Methods:

• Proposed an adaptive complementary filter for sensor fusion in hybrid HMD tracking.
• Implemented a user study in a room-scale VR environment with simulated outside-in tracking signal losses.
• Evaluated the system's performance against an Extended Kalman Filter and a state-of-the-art complementary filter.

Main Results:

• The adaptive complementary filter demonstrated improved user experience compared to the Extended Kalman Filter.
• Tracking error was reduced compared to a standard complementary filter when maintaining similar user experience quality.
• The proposed method offers configurable parameters for balancing user experience and tracking error.

Conclusions:

• The adaptive complementary filter effectively enables smooth transitions in hybrid HMD tracking systems.
• This approach enhances user comfort in VR by preventing jarring viewpoint changes.
• The method provides a flexible solution for optimizing VR tracking performance.