China is a country of vast territory with complicated geographical environment and climate conditions. With the rapid progress of the Chinese BeiDou satellite navigation system (BDS); more accurate tropospheric models must be applied to improve the accuracy of navigation and positioning. Based on the formula of the Saastamoinen and Callahan models; this study develops two single-site tropospheric models (named SAAS_S and CH_S models) for the Chinese region using radiosonde data from 2005 to 2012. We assess the two single-site tropospheric models with radiosonde data for 2013 and zenith tropospheric delay (ZTD) data from four International GNSS Service (IGS) stations and compare them to the results of the Saastamoinen and Callahan models. The experimental results show that: the mean accuracy of the SAAS_S model (bias: 0.19 cm; RMS: 3.19 cm) at all radiosonde stations is superior to those of the Saastamoinen (bias: 0.62 cm; RMS: 3.62 cm) and CH_S (bias: −0.05 cm; RMS: 3.38 cm) models. In most Chinese regions; the RMS values of the SAAS_S and CH_S models are about 0.51~2.12 cm smaller than those of their corresponding source models. The SAAS_S model exhibits a clear improvement in the accuracy over the Saastamoinen model in low latitude regions. When the SAAS_S model is replaced by the SAAS model in the positioning of GNSS; the mean accuracy of vertical direction in the China region can be improved by 1.12~1.55 cm and the accuracy of vertical direction in low latitude areas can be improved by 1.33~7.63 cm. The residuals of the SAAS_S model are closer to a normal distribution compared to those of the Saastamoinen model. Single-site tropospheric models based on the short period of the most recent data (for example 2 years) can also achieve a satisfactory accuracy. The average performance of the SAAS_S model (bias: 0.83 cm; RMS: 3.24 cm) at four IGS stations is superior to that of the Saastamoinen (bias: −0.86 cm; RMS: 3.59 cm) and CH_S (bias: 0.45 cm; RMS: 3.38 cm) models.

High-precision orbits of Low Earth Orbit (LEO) satellites are essential for many scientific applications, such as assessing the change in current global mean sea level, estimating the coefficients of gravity field, and so on. How to determinate the high-precision orbits for LEO satellites has gradually become an important research focus. HY2D is a new altimetry satellite of China, which is equipped with a Global Positioning System (GPS) and the third generations of the BeiDou Global Navigation Satellite System (BDS-3) in order to guarantee the reliability of orbital precision in radar altimetry mission. Therefore, this study adopts one month of spaceborne data to conduct the research of precise orbit determination (POD) for the HY2D satellite. Our analysis results are: (1) The standard deviation of residuals for the HY2D satellite based on spaceborne BDS and GPS data are 9.12 mm and 8.53 mm, respectively, and there are no significant systematic errors in these residuals. (2) The comparison results with Doppler Orbitography and Radio-positioning Integrated by Satellite (DORIS)-derived orbits indicate that the HY2D satellite, using spaceborne BDS and GPS data, can achieve the radial accuracy of 1.4~1.5 cm, and the mean three-dimensional (3D) accuracy are 5.3 cm and 4.3 cm, respectively, which can satisfy high-precision altimetry applications. (3) By means of satellite laser ranging (SLR), the accuracy of Global Navigation Satellite System (GNSS)-derived orbits of HY2D is approximately 3.3 cm, which reflects that the model strategies are reliable.

A two-axis stabilizing gimbal is a device that ensures a sensor is working properly on a moving platform. When classical mechanics (Newton–Euler and Lagrange) is employed to model a two-axis stable gimbal, its limitations can complicate the modeling process. To address this issue, a method for establishing a dynamic model for a two-axis stabilizing platform based on the Kane method is proposed in this paper. The Kane method offers the advantage of a simple model structure and computational efficiency. Initially, utilizing a generalized coordinate system, expressions of the generalized velocities, deflection velocities and angular velocities are derived. Subsequently, the generalized active forces and inertial forces acting on the two-axis stabilized gimbal are analyzed. Finally, by combining force and velocity with the Kane equation, the dynamic model of the two-axis stable platform is obtained, demonstrating the validity of the Kane method for establishing the two-axis stable platform model. To ensure the pointing accuracy stability of the two-axis stabilizing platform, a Novel Particle Swarm Optimization Proportion Integration Differentiation (NPSO-PID) controller is designed using the PSO algorithm. It is then simulated in MATLAB/Simulink and compared with a classical PID controller. Simulation results demonstrate that NPSO-PID exhibits superior object tracking performance compared to classical PID controllers and better optimization of control parameters compared to traditional PSO-PID controllers.

This study compares the Gravity Recovery And Climate Experiment (GRACE)/GRACE Follow-On (GFO) errors with the coseismic gravity variations generated by earthquakes above Mw8.0s that occurred during April 2002~June 2017 and evaluates the influence of monthly model errors on the coseismic signal detection. The results show that the precision of GFO monthly models is approximately 38% higher than that of the GRACE monthly model and all the detected earthquakes have signal-to-noise ratio (SNR) larger than 1.8. The study concludes that the precision of the time-variable gravity fields should be improved by at least one order in order to detect all the coseismic gravity signals of earthquakes with M ≥ 8.0. By comparing the spectral intensity distribution of the GFO stack errors and the 2019 Mw8.0 Peru earthquake, it is found that the precision of the current GFO monthly model meets the requirement to detect the coseismic signal of the earthquake. However, due to the limited time length of the observations and the interference of the hydrological signal, the coseismic signals are, in practice, difficult to extract currently.

A two-axis semi-strapdown stabilized platform is a device designed to eliminate aircraft disturbances and ensure the stability of the sensor’s orientation. A traditional two-axis semi-strapdown stabilization platform for aircraft can effectively control disturbance in pitch and yaw channel, but it cannot achieve ideal disturbance control in the roll channel. In order to solve this problem, an adaptive control method based on disturbance transformation and LWOA-PID is proposed. Disturbance transformation is the process of integrating the angular position disturbance of the roll from the previous moment into the combined disturbance of the pitch and yaw at the current moment. This is followed by decoupling the combined disturbance of the pitch and yaw at the current moment, thereby eliminating the disturbance caused by the roll from the previous moment. This process is repeated to achieve the goal of eliminating roll channel disturbances. To ensure the line of sight (LOS) pointing accuracy stability in the two-axis semi-strapdown stabilized platform system for aircraft, a whale optimization adaptive proportional–integral–derivative (LWOA-PID) controller based on Latin hypercube sampling is designed. It is then compared with the classical PID controller in Matlab/Simulink. The simulation results indicate that the disturbance conversion module proposed in this paper can eliminate the impact of roll axis disturbances on the LOS pointing accuracy of the two-axis semi-strapdown stabilized platform for aircraft. Compared to the classical PID controller, the LWOA-PID controller reduces tracking errors for step and sinusoidal signals by 50% and 75%, respectively. It also shortens optimization time by 37.5% compared to the WOA-PID while maintaining the same level of accuracy. Furthermore, when combined with the conversion module, the tracking error is reduced by an additional order of magnitude.

In UAV visual localization applications, the quality of image matching directly affects both the precision and reliability of the visual localization task. In UAV visual localization tasks, high-resolution remote sensing images are typically used as reference maps, whereas UAV-acquired aerial images serve as real-time inputs, enabling the estimation of the UAV’s spatial position through image matching. However, due to the substantial difference in imaging mechanisms and acquisition conditions between reference and real-time images, heterogeneous image pairs often contain numerous outliers, which significantly hinder the direct application of traditional matching algorithms such as RANSAC. To address these challenges, a spatial consistency-guided sampling algorithm is proposed. First, the initial correspondences are constructed based on triplet relationships, and their structural features are subsequently extracted. Then, a minimal subset sampling strategy is developed to improve sampling efficiency. Next, a data subset refinement strategy is introduced to further improve the robustness of sampling. Finally, extensive comparative experiments are conducted on the University-1652 and DenseUAV public datasets against several state-of-the-art feature matching algorithms. The experimental results demonstrate that the proposed algorithm achieves superior performance in correct matching rate, substantially enhancing the matching performance in heterogeneous image matching. Moreover, the proposed algorithm requires approximately 0.15 s per matching on average, and while maintaining the highest matching accuracy, it exhibits significantly higher computational efficiency than advanced sampling algorithms such as TRESAC and RANSAC, demonstrating strong potential for real-time applications in UAV visual localization tasks.