Autonomous navigation technology serves as an indispensable core capability for critical platforms such as unmanned systems, with its strategic value and application prospects becoming increasingly prominent. However, in GNSS-denied environments like urban canyons, indoor spaces, and underwater settings, traditional navigation methods reliant on GNSS are susceptible to interference, leading to severe accuracy degradation or positioning failure. Recent rapid advancements in deep learning technology have provided novel approaches for constructing high-precision, highly robust, and fully autonomous navigation systems. Focusing on deep learning-assisted autonomous navigation technology under GNSS-denied conditions, an in-depth review and analysis of research progress in three key areas is conducted: deep learning-assisted inertial navigation technology, multi-source intelligent navigation technology under GNSS-denied environments, and deep learning-enhanced filtering and fusion techniques. Finally, the future development trends of deep learning-assisted autonomous navigation technology are outlined.

The presence of various end-axis disturbance torques in inertial stabilization platforms limits the further improvement of the dynamic control accuracy of traditional controllers. To suppress the impact of torques on the stabilization loop, a disturbance observer is introduced into the control loop, and the dynamic performance verification is completed in a three-axis fiber-optic gyro inertial platform. Tests on a three-axis FOG platform prototype show that the disturbance observer significantly improves stabilization accuracy under dynamic conditions. The maximum misalignment angle of the three axes of the platform does not exceed 3″ under sway test conditions 6°, 1.0 Hz; 0.95°, 2.5 Hz; 0.35°, 4.0 Hz. Frequency sweep comparison tests indicate that the disturbance observer can enhance loop gain in the low-frequency band while maintaining stability margins, thereby increasing loop torque stiffness and reducing dynamic errors.

Autonomous navigation is a core capability for evaluating the robot’s level of intelligence. Traditional navigation frameworks heavily rely on continuous and precise positioning information, which often leads to system collapse in perception-degraded environments such as long corridors due to localization failure. Meanwhile, a single planning strategy is insufficient to balance efficiency and safety across diverse environments. To address these challenges, an adaptive navigation framework based on point cloud scene understanding and topological planning is proposed. A navigation strategy switching method based on SPVCNN scene understanding is developed, which effectively recognizes spatial structures such as open areas, narrow corridors, and rooms, designing an adaptive switching approach for scene-feature-oriented navigation strategies. An improved Zhang-Suen skeleton extraction method is introduced, combined with a skeleton-based pruning strategy to remove redundant nodes and branches, thereby enhancing the ability of the topological map to represent environmental spatial layouts. Furthermore, a heuristic A^* algorithm is designed, leveraging the extracted skeleton topology to generate path guidance aligned with corridor structures, improving the robot’s stability and safety margin in confined spaces. Experimental results show that, in narrow environments, the proposed method reduces navigation time by an average of 13.1% and improves average path smoothness by 34.6% compared to mainstream local path planning methods, while maintaining stable and safe operation even under localization failure.

With the continuous expansion of international trade and maritime transportation, ship trajectory prediction based on the AIS faces increasingly stringent demands for accuracy and robustness as a core technology for smart shipping and maritime supervision. To address common observational disturbances and insufficient predictive performance in complex navigation environments, a ship trajectory prediction framework combining high-precision forecasting capabilities with strong noise robustness is developed. Specifically, a novel prediction model, GRU-MSAformer, is proposed, integrating GRU and multi-scale causal self-attention mechanisms. The model first captures local temporal dependency features via GRU, then employs a multi-scale self-attention mechanism to model trajectory behavior across different time scales, thereby enabling adaptive noise filtering. Experimental results demonstrate that GRU-MSAformer achieves superior performance under both noise-free and Gaussian noise conditions. It maintains low prediction errors across 10 to 40 minutes forecasting tasks while sustaining stable prediction accuracy under varying noise intensities.

With the deep expansion of informatization into multi-dimensional physical space, dominance over spatiotemporal information has become a core area of strategic competition among major countries around the world. In this paper, addressing the development needs of the national comprehensive positioning, navigation, and timing(PNT) system, the connotation and technological evolution of dominance competition systems are explored, the cross-domain collaborative development from command of the sea, air, space to electromagnetic and information dominance are reviewed, and the technological evolution path of navigation dominance in adversarial environments is specifically examined. By summarizing the strategies and development trends of the United States, Russia, and other countries in constructing technological systems for navigation countermeasures, it conducts a critical analysis of the vulnerabilities existing in current satellite navigation systems at both the service and application levels. Furthermore, it finds out the breakthrough technological directions such as space-based resilient PNT and intelligent multi-source autonomous navigation. Finally, the future trends of navigation dominance technology from dimensions including system confrontation and the cognitive domain is prospected, aiming to provide technological support for the development of new-generation comprehensive PNT system.

Inertial measurement units play a crucial role in autonomous navigation and positioning, but their measurement errors increase exponentially over time. A pedestrian inertial navigation method based on time-frequency feature encoding neural networks is proposed. The time series sequence of inertial data and the frequency domain sequence obtained through Haar transformation are respectively used as imputs to the neural network. The time-domain and frequency-domain features are extracted separately by the inertial time-frequency feature encoder, and the dependencies between different time steps and frequency components are adaptively fused and learned through the multi-head attention mechanism. Then, the prediction results of the neural network are integrated with the inertial motion model through the extended Kalman filter framework to further optimize the state estimation. Experimental results on the public datasets TLIO and RoNIN show that, compared with the benchmark method TLIO, the proposed method reduces the ATE, RTE, and DR by 10.8%, 17.7%, and 12.9% respectively, demonstrating high accuracy and robustness in complex pedestrian motion scenarios.

Incremental angle sensors are typically used to measure the outer frame angular velocity of gyro accelerometers, making it difficult to obtain accurate outer frame position information and hindering error modeling research for gyro accelerometers. To address this, an output model of gyro accelerometer is established at small tilt angles under the gravity field. Two methods for identifying the outer frame position are proposed: one based on inner frame angle β drift and another based on characteristic points. Both methods can obtain accurate real-time outer frame position within a short power-on period and control the outer frame to halt at the target position at power-off. Test results demonstrate that the position identification and control accuracy of both methods is within ±4°. In terms of accuracy, the β-angle drift-based method outperforms the characteristic-point-based method; in terms of rapidity, the characteristic-point-based method significantly surpasses the β-angle drift-based method. The characteristic-point-based outer frame position identification method is more suitable for inertial navigation systems, meeting speed requirements while being simpler to implement. Both methods lay the foundation for error modeling and compensation of gyro accelerometers.

In complex emergency rescue scenarios such as semi-obstructed industrial sites, personnel positioning is easily affected by obstructions and signal attenuation, which makes traditional positioning solutions relying on fixed base stations difficult to be rapidly adapted to sudden rescue needs. While inertial navigation systems(INS) can provide autonomous positioning, they are afflicted with shortcomings such as cumulative errors over time and insufficient three-dimensional positioning accuracy. To address the issues of insufficient collaborative positioning accuracy and reliance on pre-existing infrastructure in emergency rescue operations within semi-obstructed industrial sites, a multi-person collaborative positioning technology combined with “ZigBee+INS” is focused on this study. The aim is to overcome the limitations of traditional positioning methods, as rapid deployment and high-precision positioning are enabled without the need for pre-established base stations, thereby rescue efficiency is improved and personnel safety is enhanced. Firstly, ZigBee anchors are deployed on the rescue personnel’s end, and the cross-power spectrum phase method is used to estimate the time delay of the time difference of arrival(TDOA). Secondly, by combining the location information of the rescue personnel and the commander, an improved TDOA algorithm is employed to suppress positioning errors. Thirdly, based on the results of the improved TDOA, the Taylor algorithm is used to determine the initial positions of the rescue personnel. Finally, position and heading constraints are established using ZigBee, barometers, and magnetometers, and multi-source information fusion and real-time position updates are achieved through an extended Kalman filter(EKF). Experimental results show that, compared with inertial navigation and classical collaborative positioning algorithms, the root mean square error and the absolute mean positioning error of the proposed algorithm are reduced by 55.42% and 62.36%, respectively. This algorithm achieves anchor-free and rapidly deployable multi-person collaborative positioning in semi-obstructed environments, and technical support is provided for precise command and control in emergency rescue operations.

The rapid development of UAV technology is gradually transforming existing operational models and giving rise to new warfare styles. In this paper, the important role played by UAV in many military conflicts and the potential security threats to countries and regions are introduced, and the difficulties and challenges faced by space-based anti-UAV combat are analyzed, the urgency and necessity of developing space-based anti-UAV systems and researching key space-based anti-UAV technologies are clarified. Subsequently, the typical types of space-based anti-UAV combat are summarized, and the key space-based anti-UAV technologies are analyzed in depth from four aspects including anti-UAV technology based on UAV, anti-UAV cluster technology based on UAV cluster, intelligent game technology, and large model technology. Finally, the research direction and development trend of future key space-based anti-UAV technology are briefly summarized from the perspective of system operation, with a view to promoting the development of future anti-UAV technology.

Traditional navigation technologies face challenges in dynamic and uncertain environments, including dependence on external signals, high energy consumption, and bulky hardware. Bio-inspired navigation technology, by mimicking biological perception and information fusion mechanisms, offers innovative solutions for efficient and robust navigation in complex environments. This paper systematically reviews the research progress in typical bio-inspired navigation sensors, categorizing them based on biological navigation modalities, with particular focus on the principles of biomimetic design, technological breakthroughs, and application potential across various sensor types. The study further summarizes current technical bottlenecks and proposes that future development should integrate brain-inspired computing with deep learning to advance the autonomous development of “perception-decision-action” full-chain systems. This work provides theoretical references and technical pathways for the engineering applications of bio-inspired navigation sensors and interdisciplinary innovation.

Axis misalignment, scale factor deviations, and time-varying noise present in low-cost IMUs significantly degrade attitude estimation accuracy. Existing neural network-based denoising methods exhibit clear limitations in multidimensional error modeling. To address this, a gyroscope adaptive calibration network integrating temporal and channel attention mechanisms is proposed: a convolutional neural network performs feature extraction, channel attention optimizes multi-axis feature weighting, and the temporal attention balances feature contributions over time, thereby enhancing network accuracy and robustness. Experimental results on the EuRoC dataset demonstrate that channel attention substantially improves dynamic compensation accuracy, while temporal attention balances accuracy across the three axes without overall performance gains, combining both mechanisms yields certain improvements, their interaction may hinder optimal performance in certain scenarios. These results validate the effectiveness of multi-attention mechanisms in inertial sensor errors modeling and provide new insights for designing low-cost gyroscope dynamic compensation algorithms.

Atomic inertial measurement systems based on the spin exchange relaxation free (SERF) principle offer advantages of high sensitivity and miniaturization but exhibit extreme sensitivity to laser power fluctuations. To address the limitations of traditional proportional integral derivative (PID) control in handling model uncertainties and low-frequency disturbances, this paper proposes a laser power stabilization method based on linear active disturbance rejection control (LADRC). For the transfer function between liquid crystal control voltage and main-path optical power established through system identification, a linear active disturbance rejection controller incorporating an extended state observer (ESO) is designed to achieve online estimation and active compensation of total disturbances. Simulation and experimental results demonstrate that this method outperforms PID control in tracking performance and disturbance rejection capability. The root mean square value is reduced from 3.07×10^－4mW to 1.17×10^－4mW, while the noise power spectral density of the gyroscope signal at 1 Hz is reduced from 1.02×10^－5(°)/s/Hz^1/2 to 7.39×10^－6(°)/s/Hz^1/2, improving stability by approximately 27.5%. This study meets the power stability requirement of SERF atomic inertial measurement systems at the level of 10^－4mW, and further provides a highly robust optimization pathway, extending the application boundaries of active disturbance rejection control in the field of quantum precision measurement.

Facing the demand for improving the efficiency of inertial navigation systems in emerging industries such as low altitude economy and unmanned systems, traditional discrete component interferometric fiber optic gyroscopes face constraints such as large volume and high cost. The new fiber optic gyroscope based on integrated optical chips has disruptive advantages in balancing gyroscope accuracy, volume, cost, and power consumption, demonstrating enormous application potential. This paper reviews the working principle and system scheme of interferometric fiber optic gyroscopes, introduces the material characteristics of thin film lithium niobate, analyzes the on-chip structure required for thin film lithium niobate chips for fiber optic gyroscopes based on its technical status, and summarizes the process steps for chip preparation. Subsequently, the relevant achievements published in the field are introduced, revealing the ability of thin film lithium niobate chips to improve system integration. Finally, in response to the difficulties and challenges encountered in the research and application process, the performance requirements for thin film lithium niobate based chips used in fiber optic gyroscopes are summarized, and prospects for future development directions are proposed.

Bionic flapping-wing aerial vehicles(FWAVs) imitate the flight mode of natural organisms(such as birds, insects), which have the advantages of strong concealment, high aerodynamic efficiency and strong ability to adapt to complex wind environment and can be used in information reconnaissance, environmental monitoring, disaster rescue and other scenarios. How to realize their autonomous navigation in complex scenes is a research hotspot in academia. Firstly, the inertial-based multi-source autonomous navigation system of the FWAV is focused on, including inertial/visual navigation, inertial/satellite navigation, inertial/visual/satellite navigation and other methods. The major advance of typical FWAV navigation systems at home and abroad is analyzed. Then, the key technologies of multi-source autonomous navigation of the FWAV are introduced. The technologies of image stabilization, perception-based localization and mapping, trajectory planning and autonomous obstacle avoidance of the FWAV are analyzed, which highlights the particularity of autonomous navigation of the FWAV and the necessity of multi-source fusion. Finally, the future development trend of multi-source autonomous navigation system of the FWAV is introduced, including open architecture, performance improvement, information fusion, group perception and other research directions.

The accuracy and reliability of geomagnetic navigation technology depend on the high-precision measurement of magnetic field information. To address the problem of inaccurate magnetic field measurement in geomagnetic navigation, a vector magnetic field real-time tracking and calibration method based on diamond nitrogen-vacancy(NV) center magnetometry is proposed. The method combines optical detection magnetic resonance technology with multi-channel microwave frequency modulation to extract the magnetic field information of each NV-axis from fluorescence signals captured by a single photodetector. Real-time demodulation and resonance frequency tracking are performed, and the voltage values are converted into three-axis magnetic fields through vector calculation, followed by calibration of the coordinate system and orthogonality. A vehicle-mounted vector magnetic field measurement experiment is conducted and compared with a fluxgate magnetometer. The three-axis errors are all less than 4%, verifying that the proposed method can perform real-time measurement of rapidly varying magnetic fields. Magnetic field changes can be detected within 10 ms, meeting the requirements for high-precision and high-sensitivity vector geomagnetic field measurement.

Quantum sensors have the potential to transcend the limitations of classical measurements. Substituting classical sensors with quantum sensors and developing high-precision quantum navigation technology emerges as a potential technological approach to boost navigation capabilities. Nevertheless, the engineering progress of quantum navigation technology has been sluggish, mainly constrained by several issues which include the complexity of quantum navigation systems, the degradation of measurement accuracy in dynamic environments, and the relatively large size and weight. In this paper, drawing on quantum navigation experiments conducted in recent years, the latest solutions to the aforementioned problems are systematically reviewed. Moreover, the most promising development directions are identified, aiming to provide valuable references for the engineering advancement of quantum navigation technology.

The system of a nuclear magnetic resonance gyroscope requires heating to 100℃~110℃ to increase the density of alkali metal atoms, thereby enhancing the spin-exchange collisional polarization effect, improving the polarization level of inert gas atoms and the signal intensity measured in the experiment. In this paper, the influence of the AC Stark effect on the measurement results for the nuclear magnetic resonance gyroscope system is considered when high-frequency current is used to heat the atomic vapor cell. A scheme is proposed to suppress the magnetic field generated by the heating coil using the interaction of multiple parallel magnetic moments in the plane. This scheme optimizes the radius distribution and current direction of each coil in space, enabling mutual cancellation of low-order magnetic field components when multiple circular coils are present, significantly reducing the residual magnetic field near the origin. Numerical calculations show that this coil design can reduce the magnetic field generated by each milliampere of current to 0.6 pT. When considering the actual heating requirements, the coil will produce a magnetic field of 25 pT, the measurement error introduced by the heating system is reduced to 10^-12 Hz, meeting the heating requirements of the nuclear magnetic resonance gyroscope while effectively improving the measurement accuracy of the system.

Achieving efficient obstacle avoidance and stable navigation in complex dynamic environments remains a critical challenge for the development of service robots. Traditional methods often rely on static maps or frequent iterative computations of locally optimal trajectories, which are prone to local optima or collisions in dynamic and narrow corridors. To address these limitations, an end-to-end navigation framework integrating spatio-temporal perception with knowledge distillation is proposed. Specifically, a PredRNN-based structure is employed to model image sequences and capture spatio-temporal features. It further incorporates a Teacher-Student architecture. The Teacher network, which utilizes LiDAR and prior knowledge of dynamic obstacles, generates high-quality policies to distill both perception and behavior into the Student network, which takes depth images as input. This enables the Student network to inherit Teacher knowledge while achieving robust decision-making. Experimental results demonstrate that the proposed method improves the success rate of reaching target destinations by 13% compared to the best-performing baseline in complex dynamic scenarios. Overall, this framework overcomes the limitations of traditional approaches in dynamic, narrow, and perception-constrained environments, exhibiting stronger generalization and adaptability, and offering a novel perspective for ensuring safe and efficient operation of service robots in real-world complex environments.

Large-scale constellation networking needs to solve the problems of orbital coordination for thousands of satellites, as well as the efficient networking of inter-satellite links and satellite-to-ground links. In the face of a vast amount of spatial data information, spatial laser communication, as a transmission method with large capacity, strong anti-interference ability, high security and fast communication rate, has become the preferred choice for spatial information transmission. As the core component of laser communication systems, the photoelectric transceiver module plays a significant role in promoting the conversion of optical and electrical signals and improving the quality of signal transmission, and it is the foundation for achieving high-speed and highly reliable laser communication. In order to meet the requirements for an optical communicator that can be used on the space satellite, a fully hermetically sealed butterfly packaged optoelectronic transceiver module is designed. Corresponding anti-radiation optimization designs are carried out in three aspects: light source, software and materials. The optoelectronic transmission performance before and after irradiation is tested, and the irradiation effect results produced by different particles are analyzed. Experiments are conducted using ⁶⁰Co γ-ray radiation and proton radiation as irradiation sources, and a measurement system for radiation field and performance parameter acquisition is built. Data comparison between the anti-radiation module and the traditional module shows that, with the increase of irradiation dose, the optical power and extinction ratio of the traditional module both decrease significantly, while the optical power of the anti-radiation module remains ≥-3 dBm, the extinction ratio still meets ≥4 dB, and the errors are all within the required range. The experimental results demonstrate the stability and good anti-radiation performance of this optoelectronic transceiver module.

Maintaining the stability of the sensitive axis of an absolute gravimeter is a prerequisite for achieving effective gravity measurement. Utilizing the stabilization based on the beam pointing control to achieve the sensitive axis direction can effectively enhance the response bandwidth of attitude of the absolute gravimeter. Regarding the control problem of the beam pointing control of the absolute gravimeter based on the quantization of the axis and the coupling with the complex optical system, a beam pointing control model based on the quantization of the axis for the absolute gravimeter is derived, and a beam pointing control alignment method is proposed. Numerical simulation analysis of the pointing control errors is carried out based on the derived control and alignment methods. The simulation results show that the average beam pointing control error introduced by measurement control noise and optical calibration error is 0.036°. Based on the numerical simulation results, control alignment experiments and beam pointing control model verification experiments are designed and carried out. The actual average error of beam pointing control measured in the control alignment and verification experiments is 0.011°, which verified the effectiveness of the beam pointing control model and alignment method.

With the widespread application of nuclear technology, the use of mobile robots to replace human operators in executing nuclear emergency tasks within unknown radiation environments has become increasingly important. However, due to limitations in detection time and sensor performance, robots can only obtain sparse radiation data. Nonetheless, in order to facilitate nuclear safety monitoring, it is essential to obtain the radiation fields distribution and the locations of radioactive sources in the environment. To address the above problems, a source localization method integrating two-dimensional laser SLAM and radiation features is proposed. This method uses mobile robots equipped with nuclear radiation detectors, LiDAR, and other sensors to collect radiation data and construct environmental maps. Subsequently, it uses Gaussian process regression method to invert the regional radiation field and integrates the inverted radiation field into the SLAM environmental map. Finally, the Hough transform method is applied to locate unknown radioactive sources. In addition, experimental verification is conducted in real environments where radioactive sources are present. The experimental results show that based on occupancy grid maps constructed using three laser SLAM algorithms (Gmapping, Hector, and Cartographer), the fusion of global radiation environment maps can be completed in both open space and factory environments, with localization accuracy exceeding 0.29 m.

In recent years, along with the development of new equipment carriers such as unmanned aerial vehicles and unmanned submarine vehicles, inertial navigation systems have generated an urgent demand for gyroscopes that combine high performance with miniaturisation, low cost and lightweight features.With the rapid development of integrated photonics, integrated optical gyroscopes have emerged, and interferometric integrated optical gyroscopes are one of the important ones.This paper outlines the basic working mechanism of IIOG based on the Sagnac effect, and provides a detailed summary of the current research status of interferometric integrated optical gyroscope from two major aspects, namely, the performance enhancement of discrete devices and the integrated coupling package of the system.The main challenges encountered in the current research and applications are analysed, and the future development trend of interferometric integrated optical gyroscopes (IIOGs) is envisioned.

The atomic spin inertia measurement system in spin exchange relaxation free (SERF) regime has the potential for ultra-high sensitivity surpass traditional inertial measurement instruments. However, the temperature error leads to the signal drifting which sufficiently decreases the inertial measurement system’s long-term stability. This paper proposes a working point optimization method to suppress the system’s sensitivity to temperature fluctuations, thereby reducing the system error caused by temperature drift. Firstly, based on the steady-state solution of the Bloch equations, the system output equation is derived, and the variation law of the system’s sensitivity to temperature fluctuations is studied. Then, an atomic spin inertia measurement device is built, and the influence of temperature working points on the temperature sensitivity and the long-term stability of the inertial measurement system is measured through experiments. The experimental results show that by using the working point optimization method, the temperature sensitivity of the system can be effectively suppressed, thereby reducing the zero-bias instability of the system by 50% and improving the long-term stability of the system.

Spin exchange relaxation free (SERF) atomic inertial measurement instruments, due to their ultra-high theoretical accuracy, have become a critical development direction for the next generation of inertial measurement devices. To address the low-frequency noise introduced by the cell temperature control process in SERF-based inertial measurement instruments, a noise feature extraction method based on variational mode decomposition (VMD) is employed. The method combines the artificial lemming algorithm (ALA) for adaptive optimization of the parameters of the VMD modal decomposition, and then quickly and accurately extracts the low-frequency interference noise features in the air chamber temperature control system. Based on the extracted noise features, the interference introduced by the data post-processing method of sliding median filtering in the temperature measurement process is localized, and an optocoupler isolation circuit based on the ACPL-C87A is designed to inhibit the reflection of electromagnetic noise by electrically isolating the power components. The experimental tests show that the designed optical isolation circuit significantly improves the device’s performance: in PID temperature control mode, the inertial measurement sensitivity at 1 Hz increases from 2.65×10^-5(°)/s/Hz^1/2 to 4.65×10^-6(°)/s/Hz^1/2, and the inertial measurement sensitivity at 2 Hz can be improved from 8.31×10^-6(°)/s/Hz^1/2 to 2.48×10^-6(°)/s/Hz^1/2. These results verify that the method effectively suppresses low-frequency noise interference, allowing the SERF inertial measurement device to maintain the fast response and high adjustment precision advantages of PID control while achieving sensitivity performance comparable to PI control.

The behavioral learning of embodied intelligence relies on high-quality robot manipulation data. However, real-world robotic data collection is costly and limited in scale, while internet-scale video data, though abundant, lacks action and state annotations. To address the challenge of extracting state representations from unlabeled videos, a video pre-training-based behavioral learning method for embodied intelligence is proposed. Firstly, an unsupervised video pre-training framework is constructed to achieve latent state extraction through feature extraction encoding, static-dynamic feature separation, and cross-frame consistency constraints. Secondly, a multimodal Transformer architecture is designed, integrating patch-wise attention mechanisms with dynamic action heads to accomplish multimodal information fusion and adaptive action generation. Simulation results demonstrate that the proposed method achieves up to 32.96% performance improvement over the baseline method Moto in task execution on CALVIN and SIMPLER simulation environments. It also exhibits significant advantages in both unknown environment generalization and environmental robustness testing, effectively enhancing the behavioral learning capabilities of embodied intelligence.

In complex underwater environments, multi-source autonomous navigation systems are often adversely affected by environmental interference, its navigation accuracy and robustness decreases. In response to the problem of degraded navigation performance in SINS/DVL/USBL integrated navigation systems under such conditions, a filtering algorithm based on a detectability quantification model is proposed. Firstly, error models are established for the SINS, DVL, and USBL subsystems, and their root mean square errors (RMSE) are calculated as quantitative indices of detectability. Then, an environmental interference factor is introduced to correct the multi-dimensional detectability model, which is constructed by considering coverage, accuracy, real-time performance, and data availability of the sensors. Based on the corrected model, an adaptive filtering algorithm is developed. Finally, the proposed algorithm is validated through test on the Yangtze River. Test results show that the proposed algorithm effectively mitigates the adverse effects of environmental interference on the SINS/DVL/USBL integrated navigation system, with RMSE reductions in the north, east, and vertical directions of 88.5%, 84.0%, and 97.4% compared to the traditional Kalman filtering algorithm, and reductions of 48.3%, 22.6%, and 18.6% compared to the maximum correntropy Kalman filtering algorithm.

To address the precise calibration challenge for high-precision inertial navigation systems, a system-level calibration method for three-autonomy inertial measurement units based on factor graph optimization is proposed. By accurately modeling high-precision IMUs, a pre-integration model is derived that takes into account Earth’s rotation, coning angular velocity, inertial device bias, scale factor, and installation error parameters. By constructing a residual model, a complete factor graph model suitable for this system is designed, and a sliding window optimization strategy is employed to iteratively solve for the calibration parameters. Simulation results show that this method is closer to the set values compared to the Kalman filter calibration method. Experimental verification using online self-calibration data from the three-autonomy IMU prototype demonstrates that the proposed method achieves performance comparable to that obtained via Kalman filter calibration. Pure inertial navigation experiments indicate that,compared with Kalman filter, parameters calibrated using this method reduce maximum eastward and northward position errors by 25.85% and 15.37%, respectively. Experimental findings confirm this method provides an effective solution for self-calibration of three-autonomy inertial measurement units, offering significant implications for expanding factor graph optimization techniques in calibration applications.

Inertial integrated navigation synthesizes information from various sensors to deliver high-precision navigation and positioning capabilities for vehicles, including aircraft, maritime vessels, automobiles and so on. In response to the challenges associated with inadequate accuracy and robustness of the SINS/GPS integrated navigation system, particularly in scenarios involving large misalignment angles and system faults, a fault-tolerant integrated navigation method utilizing the adaptive federated strong tracking unscented Kalman filter(AFSTUKF) is proposed in this paper. Firstly, an integrated navigation system integrating SINS, GPS, polarization and geomagnetic information is constructed. Secondly, the AFSTUKF is designed with a feedback reset mechanism and incorporates an adaptive fading factor along with a strategy for updating measurement noise covariance, which collectively aim to bolster estimation accuracy and robustness amidst system uncertainties and noise. Furthermore, an adaptive information sharing factor based on the Tukey function is introduced to improve the stability and overall performance of the federated filter. Experimental findings indicate an average enhancement of over 29% in estimation accuracy when compared to conventional methods. The integration of polarization and geomagnetic data contributes to an additional increase in attitude estimation accuracy of nearly 40%, thereby significantly augmenting the robustness and precision of the navigation system.

Gravity matching positioning technology is essential for positioning, navigation, and timing(PNT) systems in global navigation satellite system(GNSS)-denied environments. To further improve atomic interferometers performance in moving platforms, a high-precision attitude control method for the sensitive-axis of an atom interferometer is presented. A compact optical system, integrating a telescopic assembly with a fast-steering tip-tilt mirror, is developed to stabilize Raman laser beam alignment with respect to the atomic ensemble under angular perturbations. The designed Raman beam steering module exhibits a field of view(FOV) of ±10°, while maintaining diffraction-limited performance with a wavefront error below λ/10 across the full FOV and a beam divergence angle of less than 0.1 mrad. These characteristics ensure a position stability of the atom-laser interaction zone below 10 μm for all steering angles. The tip-tilt mirror enables precise stabilization of the Raman laser beam pointing, achieving a repeatability better than 10″ under angular disturbances of ±0.2°@100 Hz. This work provides a feasible solution to mitigate attitude errors and contrast degradation in atom interferometers operating under dynamic environments such as small unmanned surface vessels and ground vehicles.

GNSS spoofing attacks can enable attackers to covertly control autonomous vehicles, posing a serious threat to road traffic safety. Addressing limitations in existing detection methods, such as restricted application scenarios and insufficient real-time performance, a spoofing detection method based on LSTM and attention mechanisms is proposed. By constructing an LSTM-Attention model, the approach achieves multi-sensor data fusion and vehicle motion state estimation, thereby performing dead reckoning to generate a reference trajectory and employing trajectory consistency verification to detect spoofing. To mitigate the impact of cumulative errors on positioning accuracy, a sliding time window mechanism is introduced to correct errors within the window while comparing reference trajectory with GNSS navigation solutions. Experimental results demonstrate that the proposed method achieves 97.5% detection rate for abrupt spoofing attacks while maintaining low false alarm rates, outperforming existing baseline methods and meeting real-time detection requirements.

Maintenance operations in the “electromagnetic quiet zone” near radio telescopes demand high passive characteristics from navigation systems, rendering conventional navigation methods unsuitable. To address the limitations of traditional methods under passive and low-speed navigation scenarios, a polarization/geomagnetic hybrid orientation method based on BiTAN-UKF is proposed. This method first achieves preliminary nonlinear fusion of polarization navigation and geomagnetic navigation data via UKF, subsequently, by constructing a BiTAN model to capture data time-varying characteristics, it fully leverages the complementarity between the two to enhance orientation accuracy. Experimental results demonstrate that the proposed BiTAN-UKF algorithm achieves orientation accuracy of 0.375 4° in dynamic tests, providing an effective solution for passive navigation requirements in the “electromagnetic quiet zone”.

A novel guidance model is proposed for low dynamic unmanned vehicle(UV). Different from traditional guidance model whose state only contains position information, the novel model further incorporates velocity information. For the new guidance model, a new path following algorithm is designed utilizing a control approach based on Lyapunov stability theory. The algorithm consists of proportional + integral limits with saturation constraints, which is simple in form and easy to implement in engineering. Based on stability theory, it is proven that the constructed closed-loop system possesses global asymptotic stability in the presence of model uncertainties and wind field disturbances. To verify the effectiveness of the proposed algorithm, the algorithm is applied as an outer loop in the simulation of path following control of an airship, in which the simulation model is actually used in engineering. Simulation results show that the algorithm exhibits strong robustness and still demonstrates favorable path following performance even under strong wind conditions. Moreover, the guidance algorithm proposed in this paper has better convergence characteristics in simulation compared to an algorithm designed using traditional guidance model.

In autonomous driving and intelligent transportation systems, visual localization estimates vehicle pose by recognizing environmental landmarks using onboard vision sensors, while cooperative localization enhances positioning continuity and reliability in landmark-sparse scenarios through inter-vehicle information sharing. To address the challenge of global navigation satellite system(GNSS) denials, a cooperative vision-based localization analysis framework tailored for vehicular networks is proposed in this paper, and a unified stochastic geometry-based modeling and performance evaluation system is established. The spatial distribution of environmental landmarks is modeled via a homogeneous Poisson point process(HPPP), and vehicle locations are characterized by a Poisson line Cox process(PLCP). On this basis, the single-vehicle visual localization probability is derived, and a V2V-based cooperative localization mechanism is introduced when visual perception fails, incorporating Nakagami-m fading channels to obtain a closed-form expression for the cooperative localization probability. Simulation results demonstrate that the theoretical analyses match the Monte Carlo simulations with a mean square error below 1%, validating the accuracy of the proposed model. The framework reveals the coupled effects of landmark distribution, channel fading, and vehicle layout, thereby providing theoretical guidance for performance evaluation and parameter optimization of cooperative vision-based localization.

When the existing inertial measurement unit works in the space-stabilized or strapdown working mode, the sensitive axis of quartz accelerometers cannot be aligned with the thrust vector direction. This not only degrades the output accuracy of quartz accelerometers affected by the cross-coupling quadratic term error, but also prevents gyro accelerometers to take advantage of its high overload accuracy under low axial overload conditions, meanwhile affects the navigation solution accuracy. Aiming at the above problems, a thrust vector tracking mode is proposed in this paper. The classical lead-lag control method is adopted to control the radial quartz accelerometers to the zero-overload direction, ensuring that the axial accelerometer tracks the thrust vector direction. This can eliminate the quadratic term measurement error of the quartz accelerometer. Experimental verification shows that the axial accelerometer can consistently track the overload direction, and the overload sensed by the radial accelerometer can be maintained at zero under the vector tracking mode. In addition, the high-precision inertial measurement unit is equipped with a gyro accelerometer. Considering the additional output error of the gyro accelerometer introduced by the base angular motion during the thrust vector tracking control process, an output error model is established, and the gain-scheduled linear quadratic regulator (GS-LQR) is adoped to optimize the control design of the thrust vector tracking process and suppress the output error caused by the base angular motion. Simulation results show that, compared with the implemented classical lead-lag vector tracking control method, the proposed method can effectively suppress the output error of the gyro accelerometer while maintaining an equivalent tracking speed.

Rotational modulation technology is a widely applied error self-compensation technique in inertial navigation systems. Ideal rotational modulation can suppress certain device error components in the orthogonal plane of the rotation axis. However, unavoidable control errors in IMU rotation affect modulation effectiveness. On the one hand, actual rotation control suffers from errors such as angle overshoot. On the other hand, the measurement errors inherent in feedback components such as gratings within the control loop also cause non-uniform rotational angular velocity errors. Based on the consideration of controller and actuator errors, a comprehensive control system model incorporating circular grating measurement errors is established, and the unique influence mechanism of non-uniform rotational angular velocity caused by the feedback control errors of the grating on high-precision azimuth alignment is quantitatively analyzed. Calculation and simulation results show that rotational angle overshoot caused by the controller and actuator has a minor effect on the rotational modulation performance. However, when the circular grating participates in the feedback loop for angle measurement, the resulting non-uniform rotational angular velocity leads to azimuth alignment errors, which cannot be ignored in high-precision inertial navigation systems.

Redundant sensor configurations can effectively improve the reliability and navigation accuracy of inertial navigation systems(INS). Research on their calibration helps to enhance the actual operational accuracy of inertial measurement unit(IMU). Taking a decuple redundant dual-axis IMU as an example, a universal position arrangement is designed through an internal dual-axis rotation mechanism, the self-calibration consistency of the IMU in different orientations is achieved. Innovatively, the sensors are divided into three groups for system-level self-calibration. An error model for the installation of non-orthogonal sensors is established, along with navigation equations, navigation error equations, and Kalman filtering model. Based on a universal system-level calibration algorithm, a coordinate transformation module is added to calibrate all sensors. Simulation and experiment results show that both the non-orthogonal and orthogonal sensors are calibrated correctly. The calibration algorithm proposed in this paper is clear, easy to implement, and has high engineering application value.

To address the technical challenges in lightweight design for inertial measurement unit(IMU) in deep-space exploration satellites, this study breaks through the limitations of traditional aluminum alloys and innovatively adopts magnesium-lithium(Mg-Li) alloy, currently the lightest metallic structural material. By integrating structural optimization design with mechanical simulation analysis, a novel lightweight structural solution is developed, followed by systematic mechanical environmental reliability validation. Results demonstrate that the Mg-Li alloy components maintain structural integrity under 24 g random vibration and 2700 g shock loads, achieving a 30% mass reduction compared to conventional designs. To overcome the inherent corrosion resistance deficiency of Mg-Li alloys, an electroless nickel plating/painting composite coating system is implemented, significantly enhancing corrosion resistance. The successfully developed Mg-Li alloy IMU achieves a total mass of 782 g while meeting deep-space exploration mission requirements, providing a material-structure-process integrated solution for aerospace inertial device lightweight design. This work expands the application of ultra-light Mg-Li alloys in aerospace engineering.

Regarding the fact that the circularly distributed bias error of HRGs is sensitive to mechanical environments and difficult to test rapidly, the error’s mechanism and manifestation are analysed, the testing method and measurement errors based on acceleration excitation from centrifuges are discussed, and a rapid identification method for acceleration-dependent circularly distributed bias error of HRGs based on driving vibration mode precession is proposed. By driving vibration mode precession in centrifuge test, the circularly distributed bias error under different accelerations is measured without gyro rotation, and the acceleration-dependent error coefficients are obtained by fitting. Experiment results show that this method enables rapid and accurate identification of acceleration-dependent circularly distributed bias error in HRGs, effectively evaluating HRG performance accuracy.

High-precision time and frequency standards are critical for time synchronization and navigation/positioning accuracy in key fields such as global navigation satellite systems (GNSS), transportation systems, power systems, and network systems. Optical atomic clocks exhibit uncertainty metrics two orders of magnitude higher than the current primary standard, the cesium fountain clock, and their adoption for redefining the SI second has been formally proposed. The ytterbium ion optical clock, possessing two clock transition spectrums both adopted as secondary representations of the second, is one of the leading candidate ion systems for the future definition of the second. The ytterbium ion system has significant advantages, including high sensitivity to fundamental physical constants, a high clock transition quality factor, and relatively simplified laser requirements. These characteristics confer substantial potential in fundamental physics research, performance metric enhancement, and engineering applications, gaining widespread attention from research institutions worldwide. This article describes the operating principle of the ytterbium ion optical clock, reviews the key physical effects that constrain the performance of the system and the corresponding suppression methods, summarizes research progresses in fundamental physical exploration, breakthroughs in core performance metrics, and engineering technologies, and discusses prospects for its future development.

Significant viewpoint discrepancies exist between large tilt-angle UAV images and nadir view satellite maps. In the large tilt-angle UAV-satellite matching scenario, existing algorithms exhibit weak feature representation in texture-less regions, low localization precision, and insufficient robustness. To address these problems, a UAV-satellite image matching algorithm that integrates efficient channel attention with dual-stream spatial feature enhancement architecture is proposed. First, an attention mechanism is introduced into the encoder to strengthen multi-channel information interaction, thereby enhancing the network’s representation capability for blurred distant views and weak-texture regions. Second, a dual-stream spatial feature enhancement architecture is designed, which projects and fuses robust top-view features from the nadir perspective into the original oblique image via geometric transformations, achieving cross-view spatial feature enhancement. Experimental results on a multi-tilt-angle UAV-satellite image matching dataset demonstrate that the proposed algorithm achieves a localization accuracy of 57.40% within a 50 m error range, marking an 8.35% improvement over mainstream matching methods. This approach effectively resolves the problems of feature instability in weak-texture regions and information inconsistency across perspectives.