VAETT: A Velocity Adaptation-Based EKF Trajectory Tracking RFID Algorithm for HCI System

Real-time positioning using Radio Frequency Identification (RFID) faces challenges such as phase ambiguity and reduced accuracy in dynamic trajectory tracking. To address these issues, this paper proposes a Least Squares (LS) adaptive velocity based Extended Kalman Filter (EKF) trajectory tracking algorithm designed for Human-Computer Interaction (HCI) system. The algorithm first transforms the phase ambiguity problem into a distance selection problem based on the half-wavelength distance extension model. It generates an initial position candidate set using a triangular centroid weighting algorithm and introduces a dynamic threshold-weight constraint mechanism to obtain a high-quality initial position set. LS-based velocity estimates are then integrated into the EKF and updated before each antenna scheduling cycle, allowing the algorithm to generate tracking trajectories based on different initial position estimates. This approach reduces tracking errors caused by abrupt speed changes or sharp turns. To optimize computational efficiency, the algorithm dynamically discards trajectories with zero weights and selects the one with the highest weight as the final tracking result. Experimental results show that the velocity-updated EKF tracking algorithm significantly outperforms traditional EKF methods in complex trajectory tracking. The system achieves an average tracking error of 0.925 cm, improving accuracy by 3.6× and 2.26× better than RF-Pen for rectangular and circular trajectory tracking tasks, respectively.