Accurate calibration of installation errors is of paramount importance for achieving high-performance in integrated navigation systems. In this paper, a self-calibration method for addressing installation errors in an integrated inertial/polarization/celestial navigation system is proposed. The method utilizes the information from gravity, polarized E-vector, and sun to establish constrained relationship between the multiple vectors. By employing a nonlinear least squares method, the installation parameters are iteratively determined. The effectiveness of the proposed method is demonstrated through comprehensive simulation tests. The results reveal that the method achieves improved accuracy and robustness in estimating installation errors.
The configuration of spacecraft thrusters directly affects the realization of precise automatic control and propellant consumption. At present, the design of thruster configuration lacks of theoretical guidance. Spacecraft carrying out complex space missions such as rendezvous and docking are often equipped with too many thrusters. This paper proposed a theoretical design method of thruster configuration for six-dimensional position and attitude control missions. Firstly, the mathematical description of thruster configuration is introduced. Then, a general design method of thruster configuration for six-dimensional control task is proposed. Finally, simulation of relative position and attitude control of rendezvous and docking task is taken as an example to verify the correctness and validity of the above method.
In this article, the optimal attitude tracking control problem of quadrotor unmanned aerial vehicle is investigated in the presence of actuator faults and uncertain perturbation. First, based on reinforcement learning and robust control theory, the robust learning-based optimal controller is designed. Then, according to the learning-based control strategy, the fault-tolerant control is further reconstructed to suppress the influence of actuator fault. Finally, the stability analysis based on Lyapunov is given to prove the stability of the closed-loop system. Simulation results verify the practicability of the designed controller.
In the future complex and dangerous combat environment, in view of the pilot’s difficulty in processing massive/incomplete information quickly, the lack of flight control ability, the conventional flight is difficult to deal with the battlefield environment and other problems, intelligent design ideas are introduced and the pilot intelligent assistant system is proposed. This paper first introduces the development history of pilot association system (PAS) at home and abroad, and puts forward the current design requirements. Finally, it systematically presents the challenges and key technologies in the overall architecture design, decision-making technology, human-computer interaction, autonomous learning and test verification.
Aero-engine is an important component of aviation equipment. Due to the special working environment and the poor temporality of fault diagnosis, it is difficult for aircraft maintenance personnel to make accurate fault diagnosis of aero-engine. To address this problem, this paper proposes an aero-engine fault diagnosis method based on an improved snake optimization algorithm (ISO) and a bidirectional long and short-term memory network (BiLSTM). The idea of balanced pool of EO algorithm is introduced in the snake optimization algorithm to build an elite pool to compensate for the low accuracy of calculation caused by randomly selected individuals, while the Lévy flight perturbation mechanism is introduced to enhance the SO’s performance of finding the best; the optimized parameters are substituted into the BiLSTM to reconstruct the model, train the fault data for prediction and output the results. In the simulation experiments, five failure modes are diagnosed with real monitoring data of a certain type of aero-engine and compared with the BiLSTM and SO-BiLSTM fault diagnosis models. The experimental results show that the ISO-BiLSTM fault diagnosis method proposed in this paper has better fault diagnosis effect compared with BiLSTM and SO-BiLSTM, and provides a new idea for aero-engine fault diagnosis.
In light of the challenges associated with communication rejection and real-time target detection and three-dimensional perception of visual recognition systems, this paper sets out to investigate the real-time recognition and flight verification of UAV formations utilizing a monocular camera. Visual recognition systems constitute the object of study, while UAV serves as the experimental object. The research firstly focuses on devising visual formation schemes and designing system software and hardware architecture. Subsequently, considering the UAV computing power and real-time performance, a lightweight real-time target detection network is constructed to ensure target recognition speed and accuracy improvement. Relying on real-time target detection combined with ranging function, three-dimensional information of the drone is perceived. Lastly, corresponding data from UAVs is collected to train the algorithm and subsequently verify the efficacy of UAV formation and intrusion. Results indicate that the monocular visual recognition method proposed herein has both real-time detection ability and satisfactory target detection accuracy, which carries immense significance towards the development of UAVs, especially visual formation.
This paper proposes a lightweight person re-identification network that incorporates a progressive attention mechanism The network aims to address the low accuracy issue in person re-identification caused by various factors such as different viewing angles, poses, illumination conditions, occlusions, and low image resolutions. Additionally, the network design takes into consideration the need for lightweight model deployment in practical scenarios. To improve the network’s performance using limited training data, data augmentation techniques are employed to expand the training dataset and enhance the robustness of the network model. The Resnet-50 architecture serves as the backbone network, and a feature shunt structure with depthwise separable convolutions is introduced to reduce computational parameters and accelerate person retrieval inference speed. Furthermore, the feature extraction and embedding processes are separated, and a progressive attention module is introduced. This module gradually segments the features into local blocks of different granularity, allowing for the learning of discriminative features at each granularity level. This progressive approach enhances the network’s ability to perceive foreground information from coarse to fine levels and improves feature matching capability. To supervise the model, a triplet loss function is utilized, specifically designed to address challenging samples. This loss function helps reduce intra-class variations while increasing inter-class separability. The efficacy of the proposed method in person re-identification is substantiated by conducting experimental evaluation on both the Market-1501 and DukeMTMC-ReID datasets. The experimental results demonstrate that the method achieves mAP indices of 88.1% and 79.1% on the respective datasets, providing strong evidence for its effectiveness in addressing the challenges of person re-identification.
Real-time path planning of multi-UAV includes obstacle avoidance and anti-collision, which is an important condition to ensure the coordinated operation of multi-UAV. This paper studies a multi-UAV online path planning algorithm based on improved Hybird A *. Each UAV uses the Hybird A * algorithm for distributed path planning. Through the collision risk assessment of the planned path, if there is a collision risk, the path adjustment is based on the defined UAV priority level and interactive information. In this paper, the existing two-dimensional velocity obstacle model is improved, and a three-dimensional velocity obstacle model suitable for 6-DOF UAV is constructed to calculate the safe speed range of UAV. As the optimization constraint condition, the path is adjusted based on the potential field method to achieve the effect of obstacle avoidance and collision avoidance. Finally, the simulation system and experimental platform are built to verify the real-time and stability of the proposed algorithm.
At present, the aircraft fault decision-making function only deals with a single fault, but the aircraft fault has concurrency. The existing aircraft fault decision-making function lacks the ability to deal with multiple fault concurrence situations. How to trace the source of multiple fault alarm information and excavate the original fault is of great significance for simplifying the alarm display and improving the pilot’s fault handling efficiency. In this paper, a fault diagnosis and comprehensive suppression function is designed, which consists of a fast fault location method based on prior knowledge and a comprehensive diagnosis and suppression method based on fault tree knowledge. The fast fault location method based on prior knowledge is based on case reasoning, which writes the past troubleshooting cases and many elements into the fault case base. When new faults occur, the matching degree of similar cases in the case base is obtained through retrieval model, so as to quickly obtain the current fault processing method. The comprehensive diagnosis method based on fault tree knowledge converts the fault tree into a binary decision diagram, and uses Huffman coding to realize computer programming. The probability of each cut set event in the binary decision graph is the probability product of its contained bottom event, so as to determine the risk degree of the failure to locate the cause of the failure. The fault sup-pression method classifies and processes the alarm information when multiple faults occur in a single system and multiple faults occur in multiple systems. The original fault and derivative fault are filtered by using the fault correlation value, the original fault is displayed, and the corresponding derivative fault is suppressed. The fault diagnosis and comprehensive suppression function of the aircraft airborne system designed in this paper provides sup-port for the development of the large aircraft alarm system.
Electronic warfare plays an essential role in modern warfare. In this background, the multiple UAVs cooperative passive positioning technology, which has the advantages of long operating distance and strong concealment, has received significant attention. Considering the scenario of three UAVs attacking enemy surface ships in a naval battle, this paper proposes a collaborative passive localization algorithm based on TDOA and DOA to solve the positioning problem in three-dimensional space. Firstly, we establish a passive location model according to the time delay and measurement errors of the ship target and UAV. Second, the nonlinear terms are linearized using the relative spatial position between the ship and the UAV. Third, a loss function is constructed for the error term, and the least square estimation algorithm obtains the ship coordinate position. Finally, we design two comparative experiments. One discusses the influence of acute angle, isosceles, right angle, equilateral and obtuse angle five UAV spatial structures on positioning accuracy; the other explores the impact of different delay and measurement errors on positioning accuracy. The numerical simulation results effectively verify the model’s rationality and the algorithm’s effectiveness.
The purpose of this paper is to study H ∞ state feedback control issue of Takagi-Sugeno (T-S) fuzzy singular Markovian jump system (FSMJS) with constant time delays and impulsive perturbations. Under framework of linear matrix inequalities (LMIs), new criteria are derived using a modified impulse instants correlative Lyapunov-Krasovskii (L-K) functional. Through these conditions, FSMJS with constant time delays and impulsive perturbations meet H ∞ performance and achieve the stochastic admissibility. A practical suspension system is employed to illustrate feasibility of the approach.
A novel potential function multi-agent deep deterministic policy gradient (PF-MADDPG) algorithm is proposed for the multi-agent Attacker-Defender-Target (ADT). A multi-agent continuous state space and a continuous action space are established. The potential function rewards of target and defenders are designed to accelerate the game confrontation training speed, and the MADDPG algorithm is utilized to obtain effective strategies, so as to describe the influence of different actions on attackers. Finally, simulations are given to verify the effectiveness of the proposed PF-MADDPG algorithm.
Reflective material is one of the most popular materials in modern indoor and outdoor decoration because of its incredible light transmission performance, good sound insulation performance, pretty appearance and ease to clean. However, the laser sensor which is the mainstream senor used by robot to percept environment is incapable to detect transparent or reflective objects correctly. It leads that the robot can’t localize itself and generate map accurately, thus leads to serious consequence that the robot may have a collision in its working environment. There is a very real need for an approach to detect reflective material in robot working environment. This paper reviews most approaches to overcome drawback of detecting reflective material in SLAM and classifies them into three groups, namely, laser senor information only approaches, multi-sensor information-fusion approaches and artificial intelligence-based approaches. It can be a useful reference for researchers who will work in this field in the future.
Ultrasound imaging serves as an effective and noninvasive diagnostic tool commonly employed in clinical examinations. However, the presence of speckle noise invariably degrades image quality, impeding the performance of subsequent tasks, such as classification and segmentation. Existing methods for speckle noise reduction frequently induce excessive image smoothing or fail to preserve detailed information adequately. In this paper, we propose a complementary global and local knowledge network for ultrasound denoising with fine-grained refinement. Initially, the proposed architecture employs the L-CSwinTransformer as encoder to capture global information, incorporating CNN as decoder to fuse local features. We expand the resolution of the feature at different stages to extract more global information compared to the original CSwinTransformer. Subsequently, we integrate Fine-grained Refinement Block (FRB) within the skip-connection stage to further augment features. We validate our model on two public datasets, HC18 and BUSI. Experimental results demonstrate that our model can achieve competitive performance in both quantitative metrics and visual performance.
This paper mainly studies the automatic landing guidance system based on multi-scale asynchronous fusion algorithm, which can effectively improve the accuracy of landing. Firstly, on the basis of establishing the full nonlinear motion equations of six degrees freedom of carrier-based aircraft, a multi-mode guidance system composed of a variety of sensors is designed. Furthermore, in order to solve the problem of asynchronous information fusion combined with multi-mode guidance system, the adaptive unscented Kalman filter algorithm, multi-scale estimation theory and distributed federal filtering structure are adopted, and the sensors with different sampling frequencies are regarded as different scales, and the suboptimal estimation results are obtained through the local filter and sent to the main filter to obtain the global optimal estimation after fusion. Finally, the feasibility of the proposed algorithm is verified by simulation results.
Compared with the traditional anomaly detection methods, machine learning algorithms do not rely on manual and have the ability to extract advanced features of data. However, anomaly detection of spacecraft telemetry data by supervised machine learning is a challenging problem due to the lack of priori knowledge. This paper presents a signal anomaly detection algorithm based on attention mechanism. First, the long-distance characteristics of spacecraft telemetry data are captured by attention mechanism. Then, the stacked autoencoder compresses the data dimension and reconstructs the input signal to obtain the error reconstruction sequences. Furthermore, the anomaly indexes of the error reconstruction sequences are marked by the window threshold method to realize the anomaly detection of the spacecraft telemetry signal. Finally, the effectiveness of the algorithm is verified by a group of examples based on multi-channel spacecraft telemetry signals.
Global navigation satellite system plays an important role in military and civil fields, providing global positioning, navigation and timing services in real time. Unfortunately, the services provided by single constellation or double constellations cannot meet the requirements of high-level navigation, and there are still some problems such as low positioning accuracy and low number of visible satellites at specific circumstances to be solved. In this paper, multi-GNSSs are constructed and their global positioning performance is analyzed with respect to static and dynamic users. Results show that with more constellations, the position dilution of precision is decreased, and the number of visible satellites and the positioning accuracy are improved, which reveals that compared with single constellation or double constellations, multi-GNSSs can obviously improve the positioning performance. The methodology proposed can serve as a technical reference for national integrated PNT systems.
As a key mechanical component in the door system of rail vehicles, the rolling pin is closely related to the safe operation of the door system. For the purpose of maintaining the safety of the door system of rail vehicles, it is necessary to accurately predict the Remaining Useful Life (RUL) of the rolling pin. Since the degree of wear is difficult to measure, it is quite hard to predict its life in real time. Synchronously, the amount of data that can characterize the life of the rolling pin is rarely available. To predict the RUL of rolling pin online as well as provide decision support for active maintenance, this paper proposes an RUL prediction method of rolling pin based on the Convolutional Neural Network (CNN) and Bi-directional Gated Recursive Unit (BiGRU), which combines the feature extraction ability of CNN and the information retention ability of BiGRU, enabling this model to be effective in dealing with several small sample issues. The simulation results demonstrate that such a method can accurately predict the life of the rolling pin, which has essential engineering application value.
With the wide application of bus in aircraft airborne system, how to select bus efficiently is of great significance to improve the safety and reliability of flight control system. Under the background that there is no systematic method for bus selection of airborne flight control system in domestic and foreign public information, this paper proposes a systematic method for bus selection. Through the bus application of the typical aircraft type of the air-borne flight control system, the alternative bus of the airborne flight control system is preliminarily selected. For the airborne flight control system, military / civilian demand analysis, general demand analysis, special demand analysis and technical development demand analysis were carried out, and 34 demand indicators were obtained. The compliance analysis of the alternative bus is carried out according to 34 requirements indicators. If only one alter-native bus meets the above requirements, the alternative bus shall be used as the onboard flight control system bus. If there are multiple alternative buses meeting the above requirements, the choice shall be made through the analysis of the bus architecture of the flight control system.
