Head orientation prediction: delta quaternions versus quaternions

Area of Science:

• Virtual and Augmented Reality
• Robotics and Control Systems
• Human-Computer Interaction

Background:

• Display lag in head-mounted displays degrades immersion in VR/AR training simulators.
• Predictive tracking is used to mitigate display lag by anticipating head motion.

Purpose of the Study:

• To propose and evaluate a new head orientation prediction method using a Delta Quaternion (DQ)-based Extended Kalman Filter (EKF).
• To compare the performance of the DQ-EKF against a standard Quaternion EKF.

Main Methods:

• Developed a novel framework utilizing the change in quaternion (DQ) between frames to estimate head velocity.
• Employed an EKF to process the DQ for head velocity estimation and future orientation prediction.
• Tested the DQ-based framework using captured head motion data.

Main Results:

• The proposed DQ-based EKF accurately predicts head orientation.
• The DQ method achieves comparable accuracy to the computationally intensive Quaternion EKF.
• The DQ method significantly reduces the computational burden associated with head orientation prediction.

Conclusions:

• The DQ-based EKF provides an efficient and accurate solution for head orientation prediction in VR/AR simulators.
• This method enhances the value of VR/AR training by improving immersion and reducing computational costs.