Autonomous driving relies on reliable localization, but the Global Positioning System (GPS) often loses accuracy or becomes unavailable in harsh environments, such as tunnels and dense cities. Visual-Inertial Simultaneous Localization and Mapping (VI-SLAM), fusing camera and Inertial Measurement Unit (IMU) data, is a key alternative for GPS-denied scenarios. However VI-SLAM methods rarely leverage vehicles’ inherent periodic motion features, like zero velocity during stops or gear shifts, and linear motion during cruising. To address this, this paper proposes a visual-inertial localization method integrating these features into a tightly coupled VI-SLAM framework. First, based on IMU data, the Generalized Likelihood Ratio Test (GLRT) is used to detect zero velocity and linear motion. Then, Zero Velocity Update (ZUPT) residuals and Nonholonomic Constraint (NHC) residuals are constructed. These two types of residuals are combined with visual residuals, IMU pre-integration residuals, and marginalization residuals, and jointly integrated into the backend sliding-window nonlinear optimization process of Visual-Inertial Odometry (VIO). Experiments on self-collected datasets show that compared to the original algorithm, both the ZUPT integrated algorithm and the NHC residual integrated algorithm achieve higher accuracy, lower error variance, and stronger robustness.

Visual-Inertial Localization,vehicle motion feature,Zero-Velocity Update