The Global Navigation Satellite System (GNSS)/Inertial Navigation System (INS) integrated navigation technology is widely utilized in vehicle positioning. However, in complex environments such as urban canyons or tunnels, GNSS signal outages due to obstructions lead to rapid error accumulation in INS-only operation, with error growth rates reaching 10–50 m per min. To enhance positioning accuracy during GNSS outages, this paper proposes an error compensation method based on CNN-BiLSTM-Attention. When GNSS signals are available, a mapping model is established between specific force, angular velocity, speed, heading angle, and GNSS position increments. During outages, this model, combined with an improved Kalman filter, predicts pseudo-GNSS positions and their covariances in real-time to compute an aided navigation solution. The improved Kalman filter integrates Sage–Husa adaptive filtering and strong tracking Kalman filtering, dynamically estimating noise covariances to enhance robustness and address the challenge of unknown pseudo-GNSS covariances. Real-vehicle experiments conducted in a city in Jiangsu Province simulated a 120 s GNSS outage, demonstrating that the proposed method delivers a stable navigation solution with a post-convergence positioning accuracy of 0.7275 m root mean square error (RMSE), representing a 93.66% improvement over pure INS. Moreover, compared to other deep learning models (e.g., LSTM), this approach exhibits faster convergence and higher precision, offering a reliable solution for vehicle positioning in GNSS-denied scenarios.

To accurately capture the dynamic displacement of super-tall buildings under complex conditions, this study proposes a data fusion algorithm that integrates NRBO-FMD optimization with Adaptive Robust Kalman Filtering (ARKF). The NRBO-FMD method preprocesses GNSS and accelerometer data to mitigate GNSS multipath effects, unmodeled errors, and high-frequency noise in accelerometer signals. Subsequently, ARKF fuses the preprocessed data to achieve high-precision displacement reconstruction. Numerical simulations under varying noise conditions validated the algorithm’s accuracy. Field experiments conducted on the Hairong Square Building in Changchun further demonstrated its effectiveness in estimating three-dimensional dynamic displacement. Key findings are as follows: (1) The NRBO-FMD algorithm significantly reduced noise while preserving essential signal characteristics. For GNSS data, the root mean square error (RMSE) was reduced to 0.7 mm for the 100 s dataset and 1.0 mm for the 200 s dataset, with corresponding signal-to-noise ratio (SNR) improvements of 3.0 dB and 6.0 dB. For accelerometer data, the RMSE was reduced to 3.0 mm (100 s) and 6.2 mm (200 s), with a 4.1 dB SNR gain. (2) The NRBO-FMD–ARKF fusion algorithm achieved high accuracy, with RMSE values of 0.7 mm (100 s) and 1.9 mm (200 s). Consistent PESD and POSD values demonstrated the algorithm’s long-term stability and effective suppression of irregular errors. (3) The algorithm successfully fused 1 Hz GNSS data with 100 Hz accelerometer data, overcoming the limitations of single-sensor approaches. The fusion yielded an RMSE of 3.6 mm, PESD of 2.6 mm, and POSD of 4.8 mm, demonstrating both precision and robustness. Spectral analysis revealed key dynamic response frequencies ranging from 0.003 to 0.314 Hz, facilitating natural frequency identification, structural stiffness tracking, and early-stage performance assessment. This method shows potential for improving the integration of GNSS and accelerometer data in structural health monitoring. Future work will focus on real-time and predictive displacement estimation to enhance monitoring responsiveness and early-warning capabilities.