Qiangwen Fu

To improve the speed and accuracy of airborne inertial navigation systems during ground initial alignment, this paper proposes a rapid self-alignment method suitable for fixed-wing aircraft during the taxi-to-takeoff phase on the runway. This approach leverages pre-surveyed runway heading data to transform the velocity vector calculated in the navigation frame to the runway frame. By utilizing the geometric constraint that the aircraft’s actual position during ground taxiing must remain within the runway boundaries and closely follow the runway centerline, the proposed method employs a dedicated Kalman filter to achieve rapid initial alignment. Both simulations and flight tests demonstrate that the proposed method can achieve rapid initial alignment within 30 s during the runway taxiing phase, with attitude, velocity, and position accuracy levels comparable to those of GNSS-aided in-motion alignment. This method eliminates the need for external aiding equipment and can serve as an emergency initial alignment solution for airborne inertial navigation systems under GNSS outage conditions, effectively enhancing the aircraft’s autonomy and rapid response capability.

In millimeter-wave Radar-Inertial Odometry (RIO), temporal desynchronization between radar point clouds and Inertial Measurement Unit (IMU) data is a prevalent issue, primarily stemming from internal signal processing latency and the absence of external hardware triggering mechanisms. This temporal misalignment causes a severe discrepancy between observation data and state propagation in the time domain, inducing non-linear distortions in dynamic constraints and significantly degrading system robustness and localization accuracy. To address these challenges, this paper proposes a hardware-trigger-free tightly-coupled RIO framework based on Dynamics-Based Kinematic Retrodiction and Cascaded Dynamic-Geometric Filtering. First, a dynamic retrodiction mechanism is introduced to retrodict critical kinematic states (velocity and attitude)—essential for Doppler modeling—from the data arrival timestamp to the effective physical sampling timestamp, thereby achieving precise temporal alignment between system states and measurements. Building upon this, a cascaded filtering strategy is constructed, which initially employs a Global Doppler Dynamic Gate for preliminary filtering based on motion consistency, followed by RANSAC Geometric Verification to eliminate outliers induced by dynamic objects and multipath interference. Experimental results on multiple public datasets show that the proposed method achieves localization accuracy surpassing hardware-synchronized baselines without requiring hardware synchronization, while significantly enhancing system robustness and trajectory consistency in real time.

Surface defects alter the intrinsic texture characteristics, resulting in more complex and random patterns that vary significantly in scale across different types of patterns. This necessitates the use of different receptive fields. To address this challenge, a novel lightweight model is introduced. Specifically, a scale-matched block (SMBlock) is proposed, designed to adaptively adjust the desired receptive fields of stacked convolutions, thereby preventing the loss of orientation information. Additionally, heterogeneous convolution is employed to capture multi-scale features at various stages, preserving the texture details of small targets while simultaneously capturing large-scale information. Furthermore, a tripartite fusion module (TFM) is introduced to integrate features from neighboring layers, mitigating spatial misalignment. Extensive experiments on the GC10-DET and NEU-DET surface defect datasets demonstrate the effectiveness of the proposed method, achieving improvements of 3.1% and 2.9%, respectively, over the baseline. The algorithm also outperforms the baseline in detecting minimal-sized defects. The code is made available at https://github.com/HL-Louis/LSRF-YOLO.

Rotational inertial navigation system (RINS) employing error modulation technology effectively mitigates the impact of inertial measurement unit (IMU) constant drifts on positioning accuracy. However, the introduction of the rotation platform and the rotation gimbals significantly degrades the attitude measurement precision of RINS. Particularly, the attitude error caused by the non-orthogonal angles of the rotation axis markedly exceeds those caused by the IMU errors in RINS. To address this problem, a self-calibration scheme for the non-orthogonal angles of the dual-axis RINS rotation axis is proposed in this study by taking advantage of the characteristic that the attitude angles remains unchanged when the carrier remains stationary. The Kalman filter self-calibration algorithm is established by analyzing the relationship between the non-orthogonal angles and the attitude error of the carrier. Meanwhile, the self-calibration of the non-orthogonal angles of the RINS rotation axis is achieved by designing the self-calibration rotation scheme. Based on the experimental results, the fluctuation amplitudes of the pitch, roll, and azimuth angles of the carrier have been reduced by 94.07%, 95.20%, and 96.45% during static conditions through the implementation of the non-orthogonal angles calibration and compensation scheme designed in this article. Compared with traditional calibration methods, the self-calibration scheme designed in this study simplifies the calibration steps. An innovative self-calibration scheme for non-orthogonal angles is presented.

To improve the attitude measurement accuracy of the aircraft in continuous high-dynamic scenarios and ensure autonomy, a strapdown attitude algorithm based on MIMU/ADS integration is proposed. First, under the premise of missing heading information, a simplified attitude update algorithm and a carrier system velocity update algorithm are derived to reduce computational load. Next, true air speed information from the atmospheric data system (ADS) is introduced, and a MIMU/ADS velocity integration algorithm that does not depend on heading information is designed, combined with the strong tracking adaptive filter to improve the attitude accuracy. Finally, the flight test results show that, compared with the traditional horizon algorithm, this method effectively improves the aircraft’s horizontal attitude measurement accuracy in long-term, large-maneuver scenarios and offers greater environmental adaptability.

During the transfer alignment process, due to the unknown large installation angle between the master and slave inertial navigation systems (INSs), nonlinear problems may occur, seriously deteriorating the accuracy of transfer alignment. To address this issue, we propose a transfer alignment method based on graded correction of large misalignment angles. Firstly, we designed a coarse alignment algorithm based on singular value decomposition to estimate the approximate installation angle. Secondly, a virtual slave INS framework is used to convert the original large misalignment angle into a small one. Finally, the virtual slave INS is corrected using the fine alignment method, and then rotated back to the real slave INS framework. Simulation and field test results show that the proposed algorithm achieves alignment accuracy better than 0.05 °. Compared with common nonlinear transfer alignment algorithms, our method simplifies the complexity of the algorithm, significantly improves alignment accuracy and application reliability.

The current new type of inertial navigation system, including rotating inertial navigation systems and three-autonomy inertial navigation systems, has been increasingly widely applied. Benefited by the rotating mechanisms of these inertial navigation systems, alignment accuracy can be significantly enhanced by implementing IMU (Inertial Measurement Unit) rotation during the alignment process. The principle of suppressing initial alignment errors using rotational modulation technology was investigated, and the impact of various component error terms on alignment accuracy of IMU during rotation was analyzed. A corresponding error suppression scheme was designed to overcome the shortcoming of the significant scale factor error of fiber optic gyroscopes, and the research content of this paper is validated through corresponding simulations and experiments. The results indicate that the designed alignment scheme can effectively suppress the gyro scale factor error introduced by angular motion and improve alignment accuracy.

The inner lever arm effect of the accelerometer triad, also known as the size effect, means that the sensitive centers of the three accelerometers in an inertial measurement unit (IMU) cannot be mechanically installed at one point. Without compensation, it will seriously deteriorate the accuracy of calculated velocity during rotation. For the rotational inertial navigation system (RINS), this paper analyzes the influence of the inner lever arm on the rotational alignment, and proposes a multi-position alignment method based on 2-dimensional inner lever arm online estimation. In this method, the measurement center of IMU is defined at the sensitive center of a horizontal accelerometer, so that the influential inner lever arm can be simplified to two dimensions, and a fine alignment filter is designed for on-line estimation. Simulations and experiments show that the method can effectively ensure the accuracy of multi-position alignment while reducing the state dimension and calculation amount. The repeatability of the online estimation result of the inner lever arm is better than 0.1 mm.

Dual-axis rotational inertial navigation system could improve navigation performance by modulating the inertial sensor errors with rotatable gimbals. The accuracy of the system is directly affected by the initial alignment accuracy. Benefit from the dual-axis transposition mechanism, the initial alignment accuracy is improved effectively by rotating inertial measurement unit(IMU) during the alignment process. In this paper, a 6-position alignment scheme based on dual-axis rotation is proposed. The novel initial alignment scheme is designed to suppress the influence of IMU errors on initial alignment accuracy, especially the alignment accuracy of azimuth. A new initial alignment algorithm for dual-axis rotation alignment is redesigned in this paper to solve the problem of azimuth error divergence caused by traditional alignment algorithm when IMU rotates around the horizontal axis. The alignment scheme and algorithm proposed in this paper are verified by simulation and experiment.