Abstract:High-precision positioning and navigation in GPS(global positioning system) denied environments is a key technology for aircraft achieving autonomous scout, cruise, and strike. Vision navigation has the advantages of passive-type, low cost, and avoidable accumulated errors avoidable, etc. The fusion of vision and inertial navigation can give full play to their advantages and achieve the purpose of high-precision positioning. Firstly, the development of aircraft positioning technology based on multi-modal image matching assisted inertial navigation was summarized. Then this technology was elaborated in five aspects:the vision-internal calibration, multi-modal image matching, attitude algorithm, data fusion, and back-end optimization. Finally, four possible future directions were proposed as follows, two types of passive positioning combined navigation systems based on deep learning, multi-modal image matching and inertial navigation. The four possible future directions provide a reference for realizing multi-modal image matching assisted inertial navigation aircraft positioning technology.

Abstract:Accurate prediction of dynamic characteristics is an indispensable research foundation for the disturbance mechanism and vibration control of SADS(solar array drive system). The equivalent mechanical parameters of the drive mechanism were deduced with the consideration of drive control, the equivalent dynamics-characteristics analysis model of SADS was built and verified via simulation and experiment, the influence of the drive speed and control gains on the equivalent mechanical parameters of drive mechanism and dynamic characteristics of SADS was discussed. The results show that the equivalent analysis model can accurately predict the dynamic characteristics of SADS with different drive speeds and control gains, and the error between the analysis results and the test data is less than 10%. The equivalent damping of the drive mechanism cannot be affected by the drive speed and control gain, but the equivalent stiffness decreases with the decrease of the control gain and the increase of the drive speed. The drive control gradually becomes an important factor affecting the dynamic characteristics of SADS with the increase of the drive speed.

Abstract:Aiming at the complex problem of matching flight windows and orbits connection due to the multiple flight stages, long duration and complicated constraints of the manned lunar mission based on once LEO (low earth orbit) and twice LLO (low lunar orbit) RVDs (rendezvous and dockings), a design strategy of connecting the full-mission nominal flight windows and orbits with the combining layered decomposition and the forward and reverse design approach were proposed. The characteristics and application prospect of the manned lunar mission based on once LEO and twice LLO RVDs were introduced, and the basic mission requirements and engineering constraints were supposed. The effectiveness of the proposed method was verified by simulation. The nominal flight orbital elements and windows of the manned lunar mission based on once LEO and twice LLO RVDs can be responded rapidly.

Abstract:For the problem that the measurement precision of the rotation angle about line-of-sight vector of single star sensor is relatively low, a method of utilizing measurement from MEMS(micro-electro-mechanical system) gyro and two simultaneously operating star sensors to obtain accurate attitude knowledge of CubeSat was proposed. Based on the idea of averaging two quaternions, two attitude determination schemes called centralized filter and decentralized filter were drawn up by using MEKF(multiplicative extended Kalman filter). Simulation results indicate that attitude determination performance is improved effectively by the presented method and it has higher precision and quicker convergence, under the circumstance of utilizing low-cost and low-accuracy attitude sensors and traditional attitude filtering algorithm. It provides a feasible reference for low-cost and high-precision attitude determination of CubeSat, and has certain engineering application value.

Abstract:Triangle star identification has strong reliability and operability, and is still widely used at present, but the triangle star identification algorithm poses problems of redundant matching and wrong identification. The radial feature quantity of the star point was constructed by the geometric distribution of the star point. The star point was initially identified, and the navigation star corresponding to the star point was limited to several navigation stars by initial recognition. On the basis of initial identification, the triangle segmentation identification algorithm was used to identify the initial recognition result again, which improves the pertinence and efficiency of star map identification. The experimental results show that when the new algorithm is used to identify the stars, the accuracy and rapidity of star identification are improved.

Abstract:Aiming at the problem of low recognition rate of heterogeneous patterns caused by the characteristics of small sample size, strong impact, short response period and wide resonance frequency bandwidth of telemetry vibration signals, a method for sensitive feature extraction and anomaly detection of telemetry vibration signal based on referenced manifold spatial fusion learning was proposed. The multi-scale analysis method was used to decompose the signals orthogonally into each in the scale band; the multi-scale feature was extracted to construct the high-dimensional feature set. The same normal signal sample was combined with the same type of abnormal sample to establish the exclusive reference model unit, and the linear manifold learning was used to obtain the multi-scale manifold feature difference of each reference model unit to enhance the sensitivity of anomalous features. The projection matrix of each reference model unit was used to enhance the original feature set and obtain the low-dimensional multi-scale sensitive manifold feature. The input to the classifier was used to realize the state recognition of the unknown sample. The measured signal processing results verified the effectiveness of the algorithm.

Abstract:In previous studies, traditional GNSS (global navigation satellite system) ionospheric tomographic methods neglect the influence of geomagnetic field on the ionospheric variations. In this method, a new GNSS tomographic algorithm with the unequal pixel size in the height was proposed to image the ionosphere in the geomagnetic coordinate, which takes into account the influence of electron density variation at different ionospheric heights and the geomagnetic field. Also, a new iterative relaxation factor was established in the proposed GNSS tomographic algorithm to improve the accuracy of the ionospheric electron densities. The performance of GNSS ionospheric tomographic method with unequal pixel size considering geomagnetic effects was evaluated by using the IRI—2007 (international reference ionosphere 2007) model, GNSS measured data, and ionosonde data. For simulation studies, IRI—2007 was used as references, while for GNSS data, the ionosonde data were used as references. The peak ionospheric electron density error, the average absolute percentage error of electron density profile results and the root mean square error were estimated by different GNSS tomographic algorithms. The validity of the tomographic algorithm with the unequal pixel sizes considering the geomagnetic effects is verified.

Abstract:It is of great significance to accurately and quantitatively evaluate the health status of electromagnetic emission system before launch. Focusing on the large deviation of health value in evaluating the serial structure of electromagnetic emission system when applying the FAHP(fuzzy analytic hierarchy process) method, which fails to meet the requirement of nonlinear variable weight of the system, an improved FAHP-neural network health assessment method was proposed. The improved method can be calculated by constructing a nonlinear function that can satisfy the serial structure health assessment when calculating the health index of the same-level elements, and the effectiveness of the method was testified from the aspect of mathematics. On the basis of the known prior information and measured data, the neural network model was introduced to solve the nonlinear variable weight requirement of system health evaluation. An improved health evaluation model based on the pulse forming network system of an electromagnetic launch system was established and evaluation experiments were carried out. The results show that the proposed evaluation method has high assessment accuracy, and the results are in line with the actual health status of the system under various health conditions. Compared with the traditional FAHP method, the proposed method can greatly improve the accuracy of evaluation, and have no error and leakage, which verifies the feasibility and practicability of the method.

Abstract:For the promotion of submarine′s magnetic silencing ability, it is necessary to monitor the submarine′s permanent magnetic field immediately, and a reflection method of submarine′s internal and external magnetic field based on LS-SVM(least squares support vector machine) was proposed. Combined with internal and external reflection method and LS-SVM theory, an inside-out reflection model of submarine′s magnetic field was established by optimizing the model parameter with CV (cross validation). With the variation in the vertical component of the submarine′s external permanent magnetic field as an object of analysis, the extrapolation answers of simulation and hull experiment agreed well with the standard value. Compared to the RBFNN (radius basis function neural network), the proposed method has better generalization ability and extrapolation accuracy apparently, fits more in engineering facts, and can provide useful guidance in the research for closed-loop degaussing technology.

Abstract:Scientific workflow technologies in HPC are extensively applied in scientific research and engineering simulation domain. Application such as numerical simulation in complex multi-physics problems and multi-stages data process need software to compose an automatic executable workflow to increase the efficiency. There are lots of exceptions such as resource failure, task configurations errors which may cause the workflow execution to be ceased, therefore robust and continuous execution is important for workflow application. A taxonomy of fault tolerance in workflow was made and some fault tolerance techniques in typical workflow systems were reviewed. A decision-tree based event-condition-action fault tolerance model was proposed, and a non-intrusive extendable framework which was implemented in our HPC scientific workflow system HSWAP was designed. Runtime configurable error recovery strategies were also implemented in our fault tolerance software module. In order to validate our new model and framework, the fault tolerance functions were tested in real engineering simulation project. Results show that fault tolerance plays an important role in increasing workflow execution efficiency.

Abstract:The reliability of the computer system is significantly compromised by the hardware transient faults which are mainly caused by the cosmic radiation and other environmental factors. To mitigate this undesirable impact and guarantee the correctness of the running programs, a recomputation and correction mechanism for tagged instructions for an open source core named “Humming bird e203”, which is based on the RISC-V instruction set architecture, was proposed. This mechanism adds extra flag bits for each instruction and thus enables flexible recomputation for any tagged instruction at low hardware cost. Besides, it can issue the tagged instruction again automatically if the result of the first recomputation is different from the original one. This majority voting scheme can efficiently rectify most data flow errors caused by transient hardware faults. The experimental results show that with our proposal and the interrupt handler, the average probability at which programs can operate correctly can be increased by 86.67% under the random transient fault insertion.

Abstract:The research on ocean surface current sounding by along-track interferometric SAR is an important role in realizing the globalization and refinement of current information. In order to improve the iteration convergence speed and parameter retrieval accuracy of ocean surface current, the correction coefficient design problem was transformed into scale factor selection problem under constraints of the phase difference between the simulated interference phase and the measured phase difference, and radar, and platform parameter, according to the parameter characteristics of the correction coefficient to construct the constraint relation of the fitness function. The technical method was designed to calculate the correction coefficient by genetic algorithm, which was embedded in the iterative retrieval algorithm to construct a new retrieval method. The results of spaceborne SAR data simulation showed that the RMSE of current direction is better than 10.0°, and the RMSE of current velocity is better than 0.1 m/s, which meets the requirements of ocean surface current retrieval accuracy. The improved ocean surface current retrieval algorithm can reduce the number of iterations by 2 to 3, which effectively improves the retrieval efficiency. The research is of great significance to improving the effectiveness and accuracy of ocean surface current sounding by along-track interferometric SAR.

Abstract:A simplified deep learning model was proposed to solve the problem of recognition based on the strong randomness and rapid dynamic change of EEG(epilepsy electroencephalogram) signals. The proposed model utilizes one-dimensional convolutional neural network, which simplifies the convolutional layers and pooling layers to improve the efficiency. Based on the overall Keras framework, the RMSProp algorithm was used for the model in the training process, and the algorithm estimated the loss through a predefined objective function. The model design incorporated a batch normalization layer and a global mean pooling layer. The EEG recognition was researched from two aspects based on the proposed model:with empirical mode decomposition, the first three orders, the first five orders, the first seven orders, and the first eight orders of intrinsic mode functions were selected for comparative analysis on the simplified model. Because of deep learning characteristics, the proposed model can directly recognize the original EEG signals without feature extraction. After extracting 7 types of features,it adds three different methods to compare the accuracy. The experimental results show that:the recognition rate of the first three orders of intrinsic mode function reaches the level of 92.1% for the five different types of EEG signals, which is higher than that of other features. The first eight orders′ recognition rate is lower than the original signal, which indicates that data preprocessing will lead to the noise. The proposed simplified deep learning model can effectively deal with the epileptic EEG recognition problem with higher efficiency and better performance.

Abstract:Aiming at the problem of insufficient expansion ability and low recognition rate in radar emitter recognition, an intelligent recognition algorithm based on the deep learning of time-frequency feature was proposed. The shallow two-dimensional time-frequency features with high recognition and stability were quickly extracted by down sampling of short-time Fourier transform, and the noise reduction and other pre-processing were completed by using the sparseness of the local frequency-domain signal; a convolutional neural network for deep feature learning and recognition was designed, and the scale of the network was expanded by different scale convolution kernels to enhance the feature representation ability; the network was trained and tuned by using eight kinds of emitter signals under high SNR(signal-to-noise ratio) conditions, and the effectiveness of the algorithm and network was verified by a low SNR sample. The experimental results showed that the system achieves overall recognition rate of 98.31% at SNR of -8 dB, which verifies that the proposed algorithm has strong robustness.

Abstract:Aiming at the moving objects in the video sequences collected by the mobile single camera, an online multi-object automatic tracker was focused on research, which was based on the relative motion information and data association strategies. The recovery of the object trajectory was achieved by the relative motion model between the objects, and the object trajectory fragmentation was reduced. The assignment between detection of the current frame and the past trajectories was improved based on the event matching algorithm, which reduced the number of identity conversions during tracking. Experimental results show that the improved algorithm is more accurate than the original algorithm in tracking and positioning the object in the sequences, reducing the number of trajectory fragmentation and identity conversion, and our tracker achieve relative the same performance in the state-of-the-art on the TUD-Campus sequence in the international academic circle.

Abstract:Aiming at the problem that when pseudorange observation noise was high, and the ambiguity calculation of TCAR(three carrier ambiguity resolution) method was unreliable in the condition of kinematic-to-kinematic, a triple-frequency kinematic-to-kinematic ambiguity resolution with BDS/INS tightly-coupled integration was proposed. Replacing the double-difference pseudorange observations in GF(geometry-free) and GB(geometry-based) patterns with the double-difference geometry distances estimated by the tightly-coupled integration system, the method obviously decreased the noise level of the pseudorange double-difference observations and improved the success rate of triple-frequency ambiguity calculation. Simulation results show that, the high-precision position output of BDS/INS tightly-coupled integration improved the accuracy of pseudorange observations by more than 60%. In short-baseline condition, when the pseudorange observation noise is 2 m, the success rate of triple-frequency integer ambiguity calculation by GF-TCAR is 0.73% while 31.25% by GB-TCAR. However, the integer ambiguity calculation success rate by the new TCAR method is beyond 99%. And this new solution can achieve centimeter-level kinematic-to-kinematic relative positioning.

Abstract:To consider the real excitation characteristics of propeller, the measurement experiment of stern fluctuating pressure and underwater hull vibration response induced by propeller in stern wake field was conducted in circulating water channel. The experiment results show that:the fluctuating pressure has the maximum amplitude at BPF(blade-passing frequency), increases with the increase of the propeller load and decreases with the increase of the distance to propeller; the four-blade-propeller has larger amplitude at high wake region after rudders while the five-blade-propeller at low wake region between rudders; the vibration response amplitude at most of the monitor points increases with the propeller load, but there are cases that have less amplitude at BPF and no increase with propeller load; the lateral vibration induced by five-blade-propeller increases while the axis vibration decreases compared with four-blade-propeller; the vibration response amplitude at specific peak frequency presents the 1st order bending mode, and its frequency range relatively agrees with the numerical result of finite element method. The computational fluid dynamics, finite element and modal superposition method were combined to establish a numerical method to evaluate propeller induced vibration response of underwater hull. The comparison between numerical and experiment results illustrates that the numerical results provide a good agreement with the experiment results and are closer to reality than the harmonic response analysis method which uses unit harmonic excitation.

Abstract:Based on the nonlocal elasticity theory and Hamilton′s principle, the model for thermo-electro-mechanical vibration analysis was developed for a piezoelectric nanobeam embedded in viscoelastic medium. Governing equations of motion and boundary conditions for vibration analysis were derived where thermo-electro-mechanical loadings, viscoelastic foundation, nonlocal effect and piezoelectric effect were considered simultaneously. The transfer function method was employed to obtain the natural frequencies in closed form for the nanobeam with various boundary conditions. A detailed parametric study was also carried out for the effects of nonlocal parameter, boundary conditions and viscoelastic foundation on the vibration of nanobeams. The study demonstrates the accuracy and the efficiency of the developed model for vibration analysis of a complicated multi-physics system comprising nonlocal piezoelectric nanobeams, viscoelastic foundation and thermo-electro-mechanical loadings.

Abstract:For the problem of structural reliability analysis, an analysis method combining Kriging with AFOSM (Advanced first-order second moment) method was proposed. The traditional AFOSM calculation requires the gradient information of the structural limit state function. It is difficult especially when it deals with the implicit limit state function problem involving the finite element model. The proposed method combined Kriging method and AFOSM iteration fully and effectively. The Kriging method provides the gradient information of limit state function so as to overcome the difficulty of the derivation solution and improve the analysis efficiency of limit state function. Finally, numerical and engineering examples are given to verify the efficiency and feasibility of the proposed method.

Abstract:In order to effectively play the decision-making advantage of C2(command and control) organization′s centralized decision-making and collaborative decision-making, the decision-making assignment problem of Holonic-C2 organization with dynamic ability of decision-making authority was studied. In view of the shortcomings of the subjectiveness of expert fixed weights in group decision making, an expert group selection method based on the combination of expert authority and opinion consistency was proposed. This method improves the objective rationality of multi-attribute decision making. In view of the dynamic evolution of decision allocation, a decision-making evolution mechanism based on multi-stage decision making was proposed. The change of attributes in the front and back stages was considered in this method, and the decision mode transition mode in the adjacent stage was given. The simulation results show that the proposed method can give the ranking of the relatively objective decision-making modes and the multi-stage evolutionary route, which proves the feasibility and effectiveness of the method.