Abstract:Aimed to the large dispersion situation of the rocket booster or sub-stage falling to the ground without controlling, the line target homing method of parafoil system was put forward. The calculation model of rectangular boundary was established, and the contraction regional analysis of two kinds of homing methods namely point target and target line was performed. The influence of parafoil homing ability and the scattered boundary characteristics of falling area on homing performance were analyzed, and the corresponding homing strategy was obtained. A solution for setting up multiple landing areas in larger dispersion area was proposed. Finally, the dynamics and control model was established and the line target homing process was simulated. Satisfactory results are obtained，which shows that the proposed method has some reference value to the project.

Abstract:Evasion and pursuit guidance laws for the missile attacking the defense aircraft were derived based on the differential game theory. Three guidance laws which are optimalpursuit guidance law, evasion-pursuit guidance law and composite guidance law respectively were derived using a traditional cost function by defining different weight coefficients. If the missile uses the optimalpursuit guidance law, it can be intercepted easily. When the missile uses the evasionpursuit guidance law, the zero miss distance between the missile and aircraft will increase sharply; thus the missile fails to attack the aircraft easily. When the missile uses the composite guidance law, it is hard to choose the reasonable weighting parameters. To overcome these deficiencies, two new guidance laws were investigated and the sufficient conditions for the missile winning the game were analyzed. The two new guidance laws were verified by the nonlinear simulation and the results show that the proposed guidance laws can help the attacking missile evade the defender and attack the aircraft successfully.

Abstract:Aiming at improving the range of wingbody combination aircraft at hypersonic flow conditions, the aerodynamic characteristics and wing efficiency of morphing tactical missile with different deformation modes at Mach number from 3 to 8 were studied. The NavierStokes equations were used to simulate the flow field, and the lifttodrag ratio, wing efficiency, stability and controllability of different deformation modes as telescopic, variable sweep and twodimensional folding were compared. Results show that under the condition of supersonic and hypersonic flow, the variable sweep wing mode improves the lifttodrag ratio significantly, and at the same time, it has excellent wing efficiency and stability and controllability. The mode performs best at the Mach number from 3 to 8. The conclusion can offer some valuable guidance to the research and application of hypersonic morphing aircraft aerodynamic configuration.

Abstract:A novel swept flight terminal sensitive projectile was taken as a computation object, and the method to deal with simulation of the complex angular motion based on the Euler rotation theorem and a spherical sliding meshing technique was proposed. The modification value of the projectile′s angular velocity at each time step was interpolated out through applying the Rodriguez transformation matrix, and it was further allocated to the sliding meshing zone. Based on the method, the influences on pitching combination dynamic derivatives and lift coefficients under various Mach numbers made by the projectile′s rolling motion were analyzed after solving and identifying the unsteady aerodynamic parameters. Results indicate that: the proposed method can eliminate the accumulative errors in the calculation of attitude angles effectively and can realize the accurate simulation under arbitrary given angular motion; the projectile′s rolling motion generate significant influences on the identification results of pitching combination dynamic derivatives and lift coefficients, and the influence of the pitching coupled effect should be taken into account in the process of calculating the projectile′s pitching motion and stability analysis.

Abstract:With the help of the unsteady Etkin aerodynamic model, the harmonic balance method to calculate the dynamic derivatives of complicated configuration was applied. The dynamic derivatives solved by harmonic balance method agree well with the results of time domain method and experimental data, which verified the correctness of codes and the reliability of the harmonic balance method. The harmonic balance method was applied to compute the dynamic derivatives of X-51 hypersonic airframe/propulsion integrative vehicle namely WR-A. A comparison was made with the time domain method to verify the ability of the harmonic balance method for the numerical simulation of unsteady flows around the complicated model and the ability of the computing dynamic derivatives. Results show that even for the vehicle WR-A of the complex configuration, the harmonic balance method was computationally accurate and efficient.

Abstract:The numerical investigation of two main factors, velocity shear and wall proximity effect, which affects mean velocity measurement for low Reynolds number wall-bounded flow using Pitot tubes, was carried out. Their effects were separated through careful numerical simulation and cases were run for each of them respectively for error analysis. The simulation results show that: for velocity shear, the result of current work basically agrees with that of related literature, however, the nondimensional streamline shift does not asymptotically approach to a constant value; noticeable wall vicinity effect can be found within 5 tube diameters to the wall, and errors are decreasing within 1 tube diameter to the wall. Comparisons were made among the numerical solutions in this work and the corrections being used in the literature, and the new corrections based on simulation results of current work were given in the end.

Abstract:In order to fulfill the attitude control mission of micro flight vehicle delivered from a certain kind of ballistic missile, a fast attitude maneuver control approach based on the GPM (Gauss pseudospectral method) was proposed. Firstly, the accurate attitude control model was established and the coupling torque term of the reaction wheels was taken into account. Secondly, the GPM was adopted to obtain the optimal attitude trajectories. Moreover, a quasisliding mode controller was designed to track the optimal trajectories. Numerical simulation results demonstrate the optimality of the attitude maneuver trajectories obtained by using the GPM， the good performances of the quasisliding mode controller in tracking the optimal trajectories, and the good effect in rejecting the external disturbance torque.

Abstract:The existing systematic errors processing method of airborne gravity demands external gravity data, but many areas do not have external gravity data. However, semiparameter model can estimate the systematic errors without external data. Firstly, the systematic errors were modeled by using natural spline function. Then the compensation least squares method was used to estimate the parameter and the natural spline function. The smooth parameter was used to balance them. More importantly, the generalized cross validation method to determine the smooth parameter does not need prior information. Therefore, the semiparameter model was applied in the inverse Poisson integral to estimate systematic errors and downward continuation in one step. The numerical test results show that the inverse Poisson integral and least square collocation cannot estimate the systematic errors. The regularization method based on the inverse Poisson integral can reduce systematic error effect. The semiparameter combine inverse Poisson integral model can estimate the systematic errors and improve downward continuation accuracy at the same time without external gravity data.

Abstract:Scale model test is an effective approach to obtain mechanical characteristics or response of large composite spatial frame, in which the design of scale model is one of the most important parts. Large composite spatial frame was taken as the object. The limitation of the existing traditional methods was elaborated; based on the discrete similarity, the equational analysis method combing the finite element method was proposed; the similarity relationship for static response and dynamic characteristics of large composite spatial frame was deduced by this method; the 1/5 scale model of a composite keel was designed by the proposed method and was verified through experiments. The verification result shows that the proposed method and the deduced similarity relationships can be used to guide the scale model test of large composite spatial frame.

Abstract:A de-noising and recognition method based on HHT (HilbertHuang transform) was proposed to solve the problem that the traditional methods cannot effectively identify the pulse eddy current signals produced by small defects with different sizes. Firstly, the pulse eddy current signal was decomposed by EEMD (ensemble empirical mode decomposition), and the IMFs (intrinsic mode functions) of much noise were selected according to normalized autocorrelation function and its variance. Secondly, the selected IMFs of much noise were removed by the wavelet threshold denoising, and then the noiseless signal was reconstructed by adding to the nonprocessed IMFs. Then, the HMS (Hilbert marginal spectrum) was obtained by using HHT. Finally, according to the difference of HMS, the surface and subsurface defects with different sizes were identified. Experimental results show the effectiveness of the proposed method: the noise of pulsed eddy current signal is eliminated by noise elimination through EEMD, and the method based on HHT can effectively identify cracks of different sizes.

Abstract:In order to accelerate the computational speed of convolution neural network model and facilitate the implementation of largescale neural network model in embedded microprocessor, the FT-matrix2000 vector processor architecture was taken as the research background. Through the analysis of the multi-core vector processor architecture and convolution neural network algorithm, a data layout scheme was proposed in which a smaller convolution kernel data was placed in a scalar memory bank and a larger convolution matrix was placed in a vector bank. Aimed at the problem that the data in the matrix convolution is hard to reuse, a dynamic shuffling pattern with different dynamic configurable parameters based on the moving steps of the convolution kernel was proposed, by carrying out different shift operations on the convolution matrix elements, the multiplexing rate of convolution matrix data was greatly improved. Aimed at the problem that two-dimensional matrix convolution is difficult to multi-core parallelism due to the existence of data correlation, a multi-core parallel scheme with convolution matrix sharing and convolution kernel matrix multi-core exclusive was proposed. Two computing methods of convolution kernel size unchanged, convolution matrix size changed and convolution matrix size unchanged and convolution kernel size changed were designed, a performance comparison and an analysis were carried out in mainstream CPU, GPU, TI6678 and FT-matrix2000. The final experimental results show that compared with the multi-core, the CPU can be accelerated up to 238 times, compared with TI6678, the speed can be accelerated 21 times, and compared with the highperformance GPU, the speed can accelerate 663 805 times.

Abstract:For the detection of unresolved targets on the planar array radar, a rectangle array was taken as example to analyze the statistical distribution of imaginary of complex monopulse ratios in azimuth and elevation firstly. The detection model was given for planar array and an ellipse decision domain for the imaginary part of the monopulse ratios was set according to the constant false alarm theory. Besides, the real part of the monopulse ratios was used to assist the decision. By combining the two decisions, the detection algorithm for unresolved targets based on complex monopulse ratio was proposed. Simulation results based on the rectangle array for two targets demonstrate that the proposed algorithm performs well in the conditions of high signal to noise ratio, wide angle separate and suitable complex amplitude ratio.

Abstract:In order to analyze the ELF(extremely low frequency) magnetic field of ships, the corrosion related current modulated by the ship shaft was equivalent to a timeharmonic HED (horizontal electric dipole). Based on Maxwell′s equations and boundary conditions of electromagnetic field, the magnetic vector potential in air was modeled, and then the magnetic field expressions of the HED were deduced. The propagation character of the magnetic field was calculated by fast Hankel transform. The magnetic field of carbon electrodes in the pool and Tibased electrodes at sea were measured. Result shows the practicality of the magnetic field and the validity of the model.

Abstract:The equivalent problem of the code tracking bias introduced by generalized joint channel and the overlaying value of each part of the channel were researched. The influence of the code tracking bias caused by the channel characteristics of satellite-ground joint channel and each part of the channel was studied. The possible causes of pseudorange layered in the actual channel were analyzed. The nonideal characteristics of the channel affected the tracking performance of the navigation signal. The pseudo code tracking function and channel characteristics were modeled the non-ideal characteristics of both the satellite channel and the receiver channel were considered. The influence of different channel characteristics on code tracking bias was analyzed. Simulation verification was carried out with a software receiver. By analyzing the influence of several typical channels on pseudorange error, this research explains the causes of pseudorange layered phenomenon in the double-difference observation of zero-base line. Test results in practical channels prove the correctness of the analysis model. The model and analysis method can be used to explain signal quality problems and guide the optimal design of receiver channel.

Abstract:In order to solve the problem of low SNR(signal to noise ratio) radar imaging under short observation time, a compressive sensing inverse synthetic aperture radar imaging technique based on echo covariance matrix processing was proposed. The method constructs the compressive sensing problem model under the echo covariance matrix, and reduces the influence of noise on the imaging results through a specific linear transformation. By processing the simulated echo data under the condition of short observation time and low SNR, the method obtained target imaging results with higher quality and higher contrast than the traditional compressive sensing method. In the simulation experiment, the target background ratio and the background noise energy of the imaging results are better than the traditional methods, which verifies the validity of the method.

Abstract:Aimed at meshing the ferromagnetic objects with both surface elements and volume elements, a magnetic integral equation method was proposed to solve the induced magnetic field of complex-geometry carriers. Numerical results of the shell sphere and solid sphere model obtained by the proposed method can match its analytical solution well, and the ARE (average relative error) is less than 0.27%; the induced magnetic field of the ship model can be modeled well with few hybrid elements, which demonstrates the advantage of the proposed method. Based on the proposed method, the AREs between numerical and experimental results of the simple submarineequipment mockup model and the submarine mockup model are both less than 2%. Magnetism deviations caused by the equipment model are displayed by the comparison between computational results of the simple submarineequipment mockup model and the simple submarine mockup model. Thus carrier′s induced magnetic field can be calculated by the proposed method effectively, which is practically valuable.

Abstract:In order to improve robustness of the permanent magnet linear synchronous motor servo systems, an active disturbance rejection backstepping controller was presented. The unmodeled dynamics and disturbance were defined as overall disturbance and viewed as a new system state. A LESO (linear extended state observer) was designed to estimate the velocity of the mover and the overall disturbance, and convergence proof and estimation error of the LESO were given. With the linearization via dynamic compensation, a backstepping controller was designed using the outputs from the LESO. The stability of the closedloop feedback control system was given considering the estimation error of LESO. The effectiveness of the proposed controller was verified in the experimental platform, which was made by Googol Company.

Abstract:For multiple sound source localization and counting under noisy and reverberant environment, two modules, local signal-to-noise ratio tracking and coherence test, were introduced to extract the time-frequency bins less affected by noise and mutual interference of the sound sources in classical generalized cross-correlation angular spectrum; then the traditional peak search method was replaced by the dual-width matching pursuit to improve the following localization and counting accuracy. Through the simulation results, the proposed algorithm of multiple sound source localization and counting using both the filtered angular spectrum and dualwidth matching pursuit method is proved to be more accurate and robust than the traditional one, especially under the environment with lower signaltonoise ratio, stronger reverberation and more sound sources.

Abstract:In order to satisfy the requirements of position and velocity response of the laser focus control system, a three-degree-of-freedom laser focus magnetically driven micro-motion platform was designed. According to the structure of the freestyle platform, the mechanical analysis and the establishment of the kinetic equation were carried out. On this basis, the PID(proportion integration differentiation) control experiment and the robust control experiment of the fretting platform were carried out. Results show that the two control strategies can achieve a stable driving of the micromotion platform. However, under the PID control strategy, the response speed is better; while under the robust control strategy, the anti-jamming ability is better.

Abstract:In order to facilitate the design and analysis of the tracking controller, coordinate transformation was introduced to establish a typical underactuated USV mathematical model. The Serret-Frenet coordinate system was established to describe the tracking error variables. Then, considering the underactuated system, the updated law of the tracking parameters was taken as the additional input of control, the tracking control of the USV position was realized by using the improved line-of-sight guidance algorithm. In view of the disturbance such as ocean current and so on, the adaptive sliding mode technique was used to design the heading and speed tracking control law, respectively, which realizes the stabilization of heading error and longitudinal velocity tracking error. Based on Lyapunov and cascade theory, the uniformly global asymptotic and semiglobal exponential stability of the tracking control system was proved. The simulation results show that the proposed control algorithm can deal with the disturbance such as ocean currents to a certain extent, and the tracking effect is good. Moreover, it overcomes the continuous excitation of angular velocity and can track the straight and curved line at the same time.

Abstract:In order to improve the safety of linear explosive separation system, the dynamic crack propagation and crack arrest mechanism during the detonation wave propagation of linear explosive separation device were investigated on the basis of the extended finite element method and by creating 2D and 3D models. The results show that the crack propagation path is independent of the initial angle of preset crack. Without considering the loading sequence, the main directions of crack propagation are radial direction and circumferential direction for 2D and 3D models respectively. Once the loading sequence is taken into account, due to the combined effect of the stress wave from different regions, the 3D crack propagates circumferentially and axially but it does not influence the region outside the crackarrest slot either. The method used and results obtained in the study can provide a solution for the design of explosive separation devices.

Abstract:In order to probe into the design of multi-layered protective structure, the finite element method was used to simulate the typical mode and the preliquid cabin mode of the same multi-layered protective structure, and the validity of numerical simulation method was verified by comparison of the typical mode and experimental results. The influences of the pre-liquid cabin mode on the dynamic response process, the failure mode and the protective efficiency of the multi-layered protective structure were analyzed specifically, and its forming mechanism was analyzed. The differences between the two modes were compared from the following three aspects: the structure failure process, the load characteristics and the energy absorption characteristics. Finally, the protective mechanism of multi-layered protective structure is discussed, which provides a reference for the engineering design of underwater broadside protection structure of surface vessels.

Abstract:Distant supervision has been widely used for relation extraction recently. In the distant supervision, many labels may to wrongly marked, which exerts a bad impact on relation extraction. A method to reduce wrong labels was introduced by using the semantic Jaccard to measure semantic similarity between the relation phrases and the dependency terms. The training data after reducing wrong labels was used to train the relation extractors. The experimental results show that the proposed method can effectively reduce wrong labels and improve the relation extraction performance compared with the state-of-art methods.

Abstract:The previous formulas of the time complexity of impossible differential cryptanalysis proposed by Boura et al. and Derbez were researched respectively. By studying the filtration of round subkeys during the attacking procedure carefully, an improved formula estimated the real time complexity of impossible differential cryptanalysis was proposed. On this basis, the impossible differential attacks were mounted on two models of block ciphers and the time complexities by three formulas were calculated. The results show that the time complexity computed by Boura′s formula can be higher or lower than the real time complexity, and the real time complexity is 2^-1.2 times as big as the time complexity calculated by Derbez′s formula if the round subkeys are independent of each other.

Abstract:An efficient method was developed to obtain the failure probability function which combines the fractional moment-based maximum entropy method and the surrogate model method. The idea of the process is to build the failure probability function iteratively by the active learning Kriging method. Firstly, a crude failure probability function was established by using a few training samples. Then the training samples which violate the restraints of the learning function were added to update the failure probability function until the accuracy of the problem was satisfied. The fractional moment-based maximum entropy method was used to get the failure probability sample for every distribution parameter′s training sample. The samples of the failure probability could be evaluated efficiently and accurately for the optimization strategy in the fractional momentbased maximum entropy method, which could approximate the probability density function of the response effectively, and the fractional moments were estimated by the dimensional reduction method. Two examples were illustrated in the end to compare several methods such as the Bayes method, the Monte Carlo method, and so on. From the numerical results, it can be seen that the proposed method can accurately solve the problem with complex performance function and can reduce the computational cost significantly.

Abstract:For the models with interval input variables and random input variables simultaneously, a validation metric was researched to measure the agreement between the quantitative predictions from an uncertain model and relevant empirical data. With respect to different sources of uncertainty existed in engineering mathematic models and experiments，the characteristic of the validation for models with stochastic and interval mixed uncertainty was analyzed. A new validation metric for model with stochastic and interval mixed uncertainty was proposed by using the interval theory and the probability method. The properties of the proposed validation metric were discussed, and its calculation method and procedures were presented. A numerical test case and an engineering example were used to verify the feasibility and effectiveness of the proposed validation metric.

Abstract:An FPGA(field-programmable gate array) verification platform for memory link interface of manycore processor was proposed to test the function and reliability of the main components of the processor′s memory access link. Three key technologies, the on-chip read-write requests automatic generation and result checking mechanism, the hardware UDP(user datagram protocol) protocol stack in Ethernet interface and the multi-lane parallel link between FPGA chips were proposed and implemented. Experiments on the platform show that the proposed technologies are correct, they not only enrich the ways of the random request generation and result checking for functional verification, but also can test and evaluate the logics of link errors detection and lanetolane deskew. The proposed platform has been used to verify a real many-core processor in which the function of memory link interface is correct, so the validity of the verification is proved.