Nowadays the air-breathing hypersonic vehicles are focused on by the researchers of aeronautics and astronautics. This type of vehicle usually uses a scramjet engine and an integrated design method, which brings a series of aeroelastic problems. This paper first describes the research progress of integrated airframe/engine modeling of air-breathing hypersonic vehicles. Then the research status of aerothermoelastic dynamics is introduced from the aspects of aerothermoelastic/propulsion coupling, control system coupling and uncertainty analysis. The related aerothermoelastic experimental research is analyzed. Finally, several suggestions for the study of aerothermoelastic problems of air-breathing hypersonic vehicles are proposed.

The requirement of thermal comfort for passengers is constantly increasing, which puts forward more urgent requirements for the overall thermal comfort of civil aircraft cabin. Based on the actual measurement of flights flying from north to south, it is found that the temperature distribution on both sides of the cockpit is extremely uneven due to the influence of solar radiation, especially around the windows, the average temperature difference reaches 20℃, and the thermal comfort on both sides of the cockpit is quite different. Combined with the CFD dynamic simulation, based on the actual situation, the thermal comfort of the cockpit on both sides of the cockpit is quite different. Real-time measurements of temperature and thermal comfort PMV in the cabin of civil aircraft are reproduced by setting up a simulation model of equal proportion cabin. The simulation boundary condition is temperature and pressure data measured in practice. The theoretical basis for quantitative analysis of thermal comfort in the cabin of north-south flights is provided.

Selective laser melting (SLM) technique in combination with the in-situ resource utilization (ISRU) concept can be an off-world manufacturing solution to the significant engineering challenge on the large-scale construction for extra-terrestrial bases. The feasibility of SLM process applied to the additive manufacturing of lunar in-situ resource was investigated by utilizing lunar regolith simulant. A laser source was utilized to melt the powder locally in a layer-wise manner. In order to successfully fuse the powder into parts with low laser power, high efficiency and high geometrical accuracy, the SLM process parameters were investigated and evaluated by laser volume energy density. The results show that the simulant can be successfully fused into parts with high geometrical accuracy using SLM process with low laser power due to its high absorbance and low mass loss. The fabricated part quality depends on the laser volume energy density:increase of laser volume energy density input results in better mechanical properties of parts; however, excessive laser volume energy density input leads to high distortion of parts. Poor powder fluidity of the raw simulant is observed due to its complex particulate shape and a wide range of particle size distribution. The powder fluidity of the simulant is improved by optimizing its particle size range, resulting in a denser and more uniform powder bed, which can avoid defects within fabricated parts.

With the rapid development and wide application of information technology such as big data, cloud computing, Internet of Things, and mobile Internet, new combat modes continue to emerge. The "cloud operation" with the core of the task distributed command & control process becomes a brand-new cross-domain full-dimensional combat style. Based on the analysis of the feature of combat cloud and cloud operation, combined with conventional combat simulation process, cloud operation system simulation process is proposed, and the cloud operations structured simulation platform architecture design and system function design are proposed. Through cloud operation structured simulation examples, it compares the observe-orient-decide-act (OODA) cycle of traditional combat styles with cloud operation styles. The results show that the cloud operation style can effectively shorten the OODA cycle time.

A rough surface target and jamming signal recognition method for the ground frequency modulation (FM) fuze is proposed in order to effectively identify the ground target echo signal of the FM fuze and the digital radio frequency memory (DRFM) transmissive jamming. A rough surface beat frequency signal model of the ground FM fuze is established, and a two-dimensional distance-speed extraction method is used to extract the beat frequency and Doppler frequency. The beat frequency peak bandwidth and the Doppler peak bandwidth are used to identify ground target and the DRFM transmissive jamming, and their utility is verified by Monte Carlo simulation and non-parametric hypothesis statistical test. The results show that the spread of the beat frequency peak bandwidth and the Doppler frequency peak bandwidth is positively correlated with the size of the rough surface under the illumination of the fuze antenna. The Doppler frequency peak bandwidth spread is proportional to the carrier frequency. The peak bandwidth characteristics can be used to distinguish between rough surface target echo and DRFM transmissive jamming.

Aimed at the problem of limited workspace and inconsistent measurement accuracy of NDI electromagnetic tracking equipment, a method for expanding the workspace of electromagnetic tracking system and guaranteeing measurement accuracy by moving magnetic field generator is proposed. This method uses the indicator value returned by NDI system as the measurement of accuracy. When the indicator value exceeds the set threshold, the magnetic field generator connected with the manipulator is moved to relocate the sensor in the optimum working area, and the position and attitude measured by the system are unified into the coordinate system of the manipulator base through spatial transformation. In order to verify the effectiveness of the proposed method, experiments are conducted to verify that the measurement error is positively correlated with the indicator value and the distance between the sensor and the center of the magnetic field generator. Then, by comparing the errors before and after the expansion, it is shown that the mean position error can be reduced from 2.61 mm to 1.34 mm, and the mean orientation error can be reduced from 2.42° to 1.37°. This method can be used to locate and track large-scale moving instruments such as vascular interventional catheters.

In order to calculate the camera pose from a single RGB image, a deep encoder-decoder dual-stream convolutional neural network (CNN) is proposed, which can improve the accuracy of visual localization. The network first uses an encoder to extract advanced features from input images. Second, the spacialresolution is enhancedby a pose decoder.Finally, a multi-scale estimator is used to output pose parameters. Becauseof the differentperformance of position and orientation, the network adopts a dual-stream structure from the decoder to process the position and orientationseparately. To restore the spatial information, several skip connections are added to encoder-decoder architecture. The experimental results show that the accuracy of the network is obviously improved compared with the congener state-of-the-art algorithms, and the orientation accuracy of camera pose is improved dramatically.

The passive induction mechanism of electrostatic tomography (EST) determines that the number of independent measurements is equal to the number of electrodes, which is much less than the number of independent measurements of relatively mature electrical tomography (ET) technologies such as elelctrical capacitance tomography (ECT), resulting in a more severe underdetermined inverse problem. In order to address this problem, compressed sensing-based EST image reconstruction algorithm is studied. The sensitivity matrix is processed by singular value decomposition (SVD) to satisfy the restricted isometry property (RIP), and thereafter the l₁ norm regularization model and primal dual interior point algorithm (PDIPA) are utilized to reconstruct the image. Besides, constraint on the number of non-zero elements in the image vector is imposed in the iteration process according to the sparsity of debris distributed in oil. Simulation experiment demonstrates that compared to the "Circle of Appolonius" based back-projection (BP) algorithm and Landweber iteration algorithm, the aforementioned algorithm has obviously improved the imaging quality:accurate reconstruction can be obtained for single charge distributed at different positions; for two point charges whose distance is more than or equal to 1 mm, both the number and positions of the point charges can be correctly observed; for 10 groups of three randomly distributed point charge models, the accuracy rate of charged debris number monitoring is about 80%.

MEMS inertial measurement unit (MIMU) calibration is one of an important research direction in low-precision inertial navigation. Traditional calibration method has complex operating procedures and depends most on turntable accuracy. In order to overcome the problem of MIMU calibration in batch production, this paper presents a rapid micro-electro-mechanical system (MEMS) accelerometer calibration method based on adaptive genetic algorithm (GA), which converts calibration task to parameter optimization. Firstly, the principle of norm observation is adopted to establish the objective optimization function. Secondly, the optimal calibration scheme is designed on the basis of system observability analysis. Finally, calibration parameters can be optimized through adaptive GA with global search capability. Experimental results demonstrate that, compared with Newton's iteration, the proposed method can improve calibration accuracy by 1-3 orders of magnitude and increase operational speed by 61%. After the proposed calibration, the horizontal attitude error is less than 0.1° and its accuracy can reach the same order of magnitude as that in traditional method, which verifies its superiority and practicability.

The variable selection and parameter estimation problem is researched in the framework of mixed-type regression model with both functional and multivariate predictors, which broadens the scope of functional data analysis and the application fields of variable selection methodology. First the functional predictors are projected into spaces spanned by functional principal component basis functions. Then variable selection and parameter estimation are implemented simultaneously for the multivariate predictors and derived projection predictors in the form of grouping, where the tuning parameter of the penalized term is adaptively selected and the loss function is based on absolute median loss function. As to the optimization procedure, by introducing slack variables, it is transformed into a linear programming problem with several constraint conditions, which simplifies the computation. The simulation results illustrate that the proposed method performs quite well in variable selection and parameter estimation in the mixed-type regression model.

In order to realize the real-time fault diagnosis of a liquid rocket engine onboard, a fault diagnosis device is designed by combining FPGA and DSP as the hardware architecture. The FPGA controls the high-precision A/D for sensor data acquisition, the DSP runs the fault diagnosis algorithm and outputs the result. The hardware and software of the fault diagnosis device were designed separately. A recursive structure identification (RESID) algorithm is proposed for liquid rocket engine fault diagnosis. The algorithm can diagnose traffic attenuation faults in 6 ms. Based on the hardware-in-the-loop (HIL) test platform of fault diagnosis and industrial computer, the algorithm was tested and verified by the combination of automatic code generation technology and handwritten code, and observed through the upper computer interface. The results show that the RESID algorithm can accurately diagnose the common faults of the engine and realize it on the fault diagnosis device. The running time of the algorithm is 3.9 ms. The fault diagnosis device can realize real-time data monitoring and fault diagnosis, which is more compact and economical than the traditional platform. It can be used both as an onboard device and as a general platform to develop new algorithms.

In order to study the dimension reduction method of high-dimensional data based on regression model further, and the supervised clustering of variables algorithm based on Gram-Schmidt transformation (SCV-GS) is proposed. SCV-GS uses the key variables selected in turn by the variable screening idea as the clustering center, which is different from the hierarchical variable clustering around latent variables. High correlation among variables is processed based on Gram-Schmidt transformation and the clustering results are obtained. At the same time, combined with the concept of partial least squares, a new criterion for "homogeneity" is proposed to select the optimal clustering parameters. SCV-GS can not only get the variable clustering results quickly, but also identify the most relevant variable groups and in what kind of structure the variables work to influence the response variable. Simulation results show that the calculation speed is significantly improved by SCV-GS, and the estimated regression coefficients corresponding to the latent variables are consistent with the comparison method. Real data analysis shows that SCV-GS performs better in interpretation and prediction.

According to the performance requirements and technology demands of human security check in public place, combining the performance advantages of perspective imaging of passive millimeter wave imaging (PMMWI) and high discriminability in image details of visible imaging (Ⅵ), an approach for detection and localization of concealed forbidden objects on human body is presented in this paper based on the complementary advantages of PMMWI and Ⅵ. Firstly, an improved U-Net based on feature fusion in low layers is presented to enhance the sensitivity of deep neural networks (DNN) to the contour of dim small targets, and improve the accuracy of segmentation of human contours and concealed forbidden objects. Meanwhile, the pixel-level segmentation of human contours in Ⅵ is also implemented. Then, the contours of human body in PMMWI are registered with some corresponding ones in Ⅵ by scale transform and sliding fit, so the concealed forbidden objects on human body can be detected from a single frame with a high precision. Finally, the concealed forbidden objects are localized by contrasting fusion and optimizing decision according to detection results with sequence images. A series of comprehensive experiments and comparative analysis results validate the good performance of the proposed detection and localization algorithm of concealed forbidden objects on human body towards security check of public places.

To solve the problem that the hole mapping method occupies too much physical memory, an improved hole mapping method was developed. Based on the neighbor-to-neighbor search algorithm, a donor search method based on adjacent front was developed. An explicit assembly algorithm of unstructured overset grid was presented by combining the cut-paste method with the implicit cutting technique. First, the algorithm generated a set of Cartesian grids surrounding the wall surface. Second, those Cartesian cells intersecting the wall surface were stored. Finally, relative positions of the stored Cartesian cells were used to determine whether a grid point was inside the wall. After successfully determining all the grid points inside the wall, the current fringe grid points were used as the initial front, and the overlapping area was optimized by the wall distance of each grid point to generate the final interpolation boundary. The proposed explicit algorithm optimizes the traditional implicit assembly process of unstructured overset grid. It features in low physical memory occupation, low cost of donor searching and high computational efficiency. The accuracy and applicability of the proposed explicit method were verified by two typical complex flow examples.

Array jet impingement cooling technology can effectively solve the heat dissipation problem of high heat flux devices. In order to verify the effectiveness of heat transfer enhancement on the impacted surface for optimizing cooling performance of two-phase jet cooling, this article studied the effects of different pin-finned surface structures on the flow and heat transfer characteristics of confined array jet cooling combined with high-speed microscopic imaging methods. Two kinds of pin-finned surface morphology were designed:smooth cutting needle rib (0.6 mm×0.6 mm×1.0 mm) and rough needle rib with porous sintered layer (particle size 73~53 μm). In the experiment, jet cooling heat sink with smooth surface was used as the control group, anhydrous ethanol was used as the working medium, and all the inlet temperatures were the same (20℃). When the flow rate is 7.5 mL/s and the heating heat flux increases from 5 W/cm² to 100 W/cm², the heat transfer coefficient of the heat sink continues to increase but the increase rate gradually decreases, and no phase change is observed. Under the experimental conditions of changing the fluid flow rate (fluid Reynold number) with fixed heat flux 82.6 W/cm², 80.5 W/cm², when the flow rate decreases from 7.5 mL/s to 1.0mL/s, it can be clearly observed that the working fluid in the jet cavity gradually enters bubble flow, slug flow and annular flow from stratified turbulence flow, which correspond to the initial boiling zone, nuclear boiling zone and membrane boiling zone, respectively.

With the development of digital signal processing technology, the application of high-performance signal processing has attracted more and more attention, which also poses great challenges to the computing speed and throughput efficiency of the corresponding processors. The shifter unit is an important component on the digital signal processor (DSP). By designing additional dedicated random access memory (RAM) and look-up table (LUT) for the shifter unit, this paper optimizes and adjusts its instruction set and architecture, so as to improve the use efficiency and transmission rate of the processor. In addition, based on the shifter and the corresponding look-up table instruction, it can carry out shift, extraction, arithmetic and logical operation processing at the same time of data temporary storage. And the process of the partial data operation is directly merged into the data read/write process of the shifter RAM, which greatly improves the efficiency of arithmetic unit. Experiments show that the temporary storage technology based on the shifter look-up table can achieve the throughput rate close to the transmission bus, and the signal processing algorithm fast Fourier transformation (FFT) can achieve the performance improvement of the acceleration ratio of 1.15 to 1.20.

For the harsh internal signal environment and the large signal attenuation of satellite navigation with the difficulties of capturing and tracking signals stably inside large-scale, complex and multifunctional buildings, the multipath effect is serious, and the short multipath has a great influence on the positioning accuracy. Taken in this sense, it is difficult to locate directly with the navigation signals. This paper proposes a method of indoor positioning method based on a large area of ground base station (building group). According to the relationship between frequency and signal penetration performance and spatial attenuation, the frequency band of very high frequency is selected as the signal carrier, with the combination of the code pseudorange and the carrier pseudorange, to achieve the coverage and positioning performance. In this paper, a basic test system is built by using the proposed new positioning method. The signal acquisition, tracking and pseudorange differential are carried out through signal generation, transmission and wireless propagation, which preliminarily verifies the feasibility and the coverage ability of the proposed method.

In order to analyze the reasons why the design of forward-swept wing aerodynamic configuration cannot be popularized and applied in aviation industry, simplified strake-wing, canard-wing and strake/canard-wing configurations were constituted by fixing strake and canard on forward-swept wing and back-swept wing, so as to deeply understand the flow characteristics and the mechanism of vortices interference between the two different configurations of forward-swept wing and back-swept wing. and First, the reliability and accuracy of numerical computation method were validated by comparing the computing results with experimental data of a standard model. Then, the lift coefficient curves of different configurations were obtained through numerical computation. Finally, the complex vortex interaction mechanism of different configurations were analyzed by pressure contours and streamlines. The results indicate that induction and convolution between vortexes of configurations based on back-swept wing enhance the lift coefficient and increase stalling angle of attack, and the effect was more apparent on the configuration fixed with strake and canard. There is no convolution effect between vortices of configurations based on forward-swept wing, and vortexes of configurations based on forward-swept wing perform an adverse interaction by bumping and squeezing, which makes the lift coefficient of the forward-swept wing much lower than that of the back-swept wing at high angles of attack. The leading-edge vortices of the forward-swept wing cannot be coupled with the canard wing vortices and strake vortices at high angles of attack, and cannot make full use of the non-linear lift force, which is the shortcoming in the aerodynamic layout design of the forward-swept wing.

Aimed at the problem that the signal of frequency-modulated continuous wave (FMCW) fuze is very difficult to detect at ultra-low signal to noise ratio (SNR), a detection system based on small-scale periodic state Duffing oscillator is established. This system combines Duffing oscillator characteristics with stopping oscillation system theory, eliminating the inherent deficiencies of the traditional transformation-dependent Duffing oscillator detection methods, extending the frequency detection range through a single Duffing oscillator, and reducing the computing cost. On this basis, the phase trajectory characteristics of the small-scale periodic state are analyzed, and then a FMCW fuze signal detection method based on small-scale periodic state Duffing oscillator is proposed. The experimental results show that the small-scale periodic state Duffing oscillator detection method has an average detection error of less than 1% for the real FMCW fuze radiation signal at -30 dB ultra-low SNR, which verifies the effectiveness of the proposed method.

To meet the requirement of reference atmospheric model in aerospace application, that is, the model should have global scale coverage across seasons and space, and high accuracy of key trajectory points, such as the range for takeoff and landing, a method for modeling and application of range reference atmospheric density is proposed. First, the characteristics of atmospheric density in typical months of winter and summer, including monthly mean and density perturbation, are analyzed. On this basis, a quantitative correction method of the global reference atmosphere model (GRAM) based on sounding measurement is proposed in this paper. Then, the range reference atmospheric density model with atmospheric perturbation and seasonal variation is constructed. Finally, the transition method and application method for the transition from the range reference atmospheric density model to GRAM are proposed. The simulation results show that the atmospheric density characteristics of the measured area have obvious seasonal differences in winter and summer, and the atmospheric density model should be built according to the season. The results of Monte Carlo simulation of the range reference atmospheric model show that it can effectively simulate the atmospheric density characteristics of the measured data. Combined with GRAM transition, the reference atmosphere model has both global coverage and high range accuracy.

Single shot multibox detector (SSD) owns the relatively independent regression computations of multi-regressive feature maps, while the object detection algorithms based on SSD cannot make a tradeoff between detection accuracy and real-time speed. To solve the problems above, a single shot mutibox detector based on asynchronous convolution factorization and shunt structure (FA-SSD) is introduced based on asynchronous convolution factorization algorithm and shunt structure. The shunt structure, based on the proposed asynchronous convolution factorization algorithm, is designed to staggerly connect the layers of regression features, enhancing the unity and coordination between regression calculations. In order to optimize the mainstream of high-level structure, the asynchronous convolution factorization algorithm and max pooling are implemented to reduce the dimension of image features in the mainstream and shunt respectively, which can hold the spatial information while improving the diversity of features. According to the experimental results from VOC2007test, FA-SSD achieves a mean average precision of 80.5% after the training of VOC2007trainval and VOC2012trainval with nominal resolution of 300×300, while the detection speed exceeds 30 frames per second.

In order to realize the registration of point cloud data respectively obtained from LiDAR and camera, we used a fast multi-scale registration (FMSR) algorithm to register the point cloud data, based on a self-designed three-dimensional scanning laser radar system in our laboratory. The algorithm includes two steps:coarse registration and fine registration. In the coarse registration, an adaptive scale key point quality (ASKQ) algorithm was used to match key points and determine the initial parameters for fine registration. And in the fine registration, K-nearest neighbors (KNN) algorithm was used to simplify the search process and improve the algorithm efficiency. The optimal rotation matrix, translation vector and scale factor between two sets of point cloud data were obtained through many iterations. The simulation verified the stability of FMSR algorithm for multiscale registration. Simulation and experimental results show that the proposed algorithm successfully registers the point cloud data of the self-made LiDAR system and commercial camera. The root-mean-square error of the registration is 0.194 m and the execution time is 16.207 s, for a building with size of 20.30 m×7.85 m×26.56 m. Compared with an existing scale-iterative closest point (S-ICP) algorithm, the registration accuracy of the proposed algorithm is improved by 0.131 m, and the execution time is reduced by 30%. The proposed point cloud registration method can provide an algorithm basis for scene reconstruction and texture matching.

The unmanned aerial vehicle (UAV) manipulator's flying performance mostly relies on experts' subjective evaluation and different flight maneuvers lack pertinent evaluation criteria, so a model which uses flight data to evaluate the horizontal flight quality of UAV is established. Firstly, the flight data segment of UAV's horizontal flight maneuver was identified by the flight discrimination rules. Then, according to Bollinger bands theory, the scores of multiple flight parameters in each flight data segment were calculated. Finally, the weight of each parameter was determined by the entropy weight method and the indexes reflecting the horizontal flight maneuver's quality of different UAV manipulators were obtained. In a quadrilateral flying training mission, four groups of different UAV maneuvers' flight data and one group of flight data under autonomous control were input into the model. The calculation results show that the model can well identify the horizontal flight maneuver and distinguish the horizontal flight maneuver's quality of different manipulators, which can provide advice for the training of UAV manipulators.

The interacting multiple model (IMM) estimator has been proven to be of excellent performance and low complexity in tracking agile targets. The success of IMM attributes to mode mixing, where model outputs are mixed for model-conditional reinitialization. The problem of unequal dimension states mixing in IMM estimation is studied and an optimal method for IMM mixing is proposed based on summarizing the existing methods. By introducing the concept of "switching" state into the target state, the new method dynamically adjusts the hybrid strategy with model probability and innovation to achieve optimal estimation. The simulation results show that the proposed approach outperforms the existing algorithms in the scenarios of mixing different models.