Visual–Inertial Navigation System Based on Virtual Inertial Sensors

In this paper, we propose a novel visual–inertial simultaneous localization and mapping (SLAM) method for intelligent navigation systems that aims to overcome the challenges posed by dynamic or large-scale outdoor environments. Our approach constructs system utilizing virtual inertial sensor components are mapped torso IMU under different gait patterns through classification. We apply zero-velocity update (ZUPT) initialize with original information. The pose information is then iteratively updated nonlinear least squares optimization, incorporating additional constraints from ZUPT improve accuracy of system’s positioning capabilities in degenerate Finally, corrected fed into solution. evaluate performance our proposed SLAM three typical environments, demonstrating its applicability high precision across various scenarios. represents significant advancement field offers promising solution