17 October 2024 in Measurement Science and Technology
To improve the alignment accuracy and environmental applicability of the micro-electromechanical system (MEMS)-based strapdown inertial navigation system (SINS), this study proposes a global navigation satellite system (GNSS)-assisted multi-vector determination of the attitude optimal indirect coarse alignment. Using the inertial information output from the inertial measurement unit and the velocity information from the GNSS, a simplified velocity observation vector is constructed and the velocity lever arm stemming from the mounting position inconsistencies of the GNSS and SINS is fed back into the velocity of the carrier system. When constructing the observation vectors, the integration interval is shortened by reconstructing the two-vector integration formula for reducing the cumulative error of the inertial device. The attitude matrix problem is defined as the Wahba problem, which is solved using the singular value decomposition method. Based on the relationship between gyro zero bias and misalignment angle, the corresponding state and measurement equations are designed. Furthermore, owing to the measurement noise uncertainty, the adaptive traceless Kalman filtering algorithm is introduced to realize the effective adaptation processing of the measurement noise. More accurate attitude matrix estimates are obtained by continuously correcting the carrier system transformation matrix. The running car experiment results show that the proposed method exhibits higher alignment accuracy and environmental applicability than the current MEMS strapdown inertial navigation coarse alignment method and the traditional optimization-based alignment method.
1 November 2026 in Engineering Research Express
Traditional visual simultaneous localization and mapping (SLAM) methods generally rely on the static-scene assumption and may suffer substantial localization degradation in the presence of moving objects. To address this issue, this paper proposes a dynamic visual SLAM method based on semantic and scene-flow constraints and integrates it into the ORB-SLAM3 framework. First, YOLO11s is used to detect potentially dynamic objects and construct semantic prior regions. Second, adjacent-frame depth information and relative camera motion are combined to build a dynamic-feature discrimination model based on depth-normalized scene-flow residuals and deviations from a global motion reference. Finally, a dual-threshold strategy applies different criteria inside and outside the semantic prior regions, retaining only static features for pose estimation and back-end optimization. Experimental results show that, in terms of absolute trajectory error root mean square error, the proposed method reduces the error by an average of 69.20% on the TUM indoor dynamic sequences and 17.87% on the KITTI outdoor road sequences relative to ORB-SLAM3, while also achieving lower errors than representative dynamic SLAM methods on the evaluated sequences.
3 November 2026 in Measurement Science and Technology
Integrated navigation systems often have limited awareness of environmental changes in complex settings. When global navigation satellite system (GNSS) signals are disrupted, mismatch in the measurement-noise model can substantially degrade positioning accuracy and stability. To address this problem, we propose an interacting multiple-model-driven innovation-adaptive factor graph optimization method (IMM-IAFGO). First, the IMM framework runs multiple noise models in parallel and updates their posterior probabilities using measurement innovations and prior knowledge, thereby estimating the measurement-noise covariance. The resulting estimate is then used to adaptively reconstruct the observation-information weights of the GNSS factor nodes in the factor graph. Second, robust estimation is achieved through a dynamic adjustment mechanism based on inter-frame innovations. Under severe GNSS signal distortion, odometry (ODO) aided by inertial measurement unit attitude feedforward is introduced to constrain the inter-frame displacement increments. Real-world vehicle experiments demonstrate that the proposed method significantly outperforms error-state Kalman filter, factor graph optimization, and adaptive factor graph. Specifically, it reduces the horizontal positioning root mean square error by 87.92%, 82.85%, and 23.72%, and the mean absolute error by 90.68%, 82.33%, and 26.38%, respectively. This method provides an effective strategy for achieving continuous and high-precision positioning in vehicle integrated navigation systems operating in complex environments.