Abstract:Taking the fluctuating pressure at the outlet of the engine compressor as a surge detection signal, the statistical characteristics of the fluctuating pressure at the outlet of the compressor being tested and analyzed, and the fluctuating pressure signal being transformed normally by the Johnson transform method of sliding window based on the statistical characteristics, a threshold and a method for surge detection adapted to any state of the engine are proposed respectively. The method is successfully used to detect the surge failures at a certain engine for three times. The analysis shows that the fluctuating pressure at the outlet of the engine compressor does not obey the normal distribution, and the success rate of the normal transformation for the fluctuating pressure in the steady state process is higher than that in the transient process, the sample capacity reduced can effectively increase the success rate of normal transformation in transient process. For the upper and lower boundary distances of the 99.95% probability distribution of the fluctuating pressure, the amplitude range varies greatly at different speeds and increases with the increase of the speed. According to the proposed surge detection amount, a fixed detection threshold can be set adaptable to any state of the engine, and the detection time of starting engine surge and exiting engine surge is less than 20 ms.

Abstract:Aimed at the problems that in the traditional gear fault signal processing, the sensitive characteristic information is easily lost in the process of the single signal processing , a formal sampling wavelet (MUDW) fusion pretreatment method of multipoint gear fault vibration signal is proposed to make full use of the characteristic information of each signal contained to reduce the influence of noise and interference elements. First of all, each channel signal is preprocessed by using the time synchronous average (TSA), and then the morphological undecimated wavelet framework signals are decomposed, and each decomposition level approximate signals contribution to the fault is characterized to raise the proportion of useful approximation signals by adopting the correlated kurtosis (CK) and the weighted operation form. On the basis of this, each decomposition level signal is fused by the established fusion rule, improving the reconstructed 〖JP2〗signal characteristic information. Finally, the experiment of gear crack fault shows that the method can restrain noise well, and the meshing frequency and its double frequency are prominent. The fusion effect is ideal.

Abstract:Taking a micro turbine engine as the platform, numerical simulation and experimental verification of convergent axisymmetric vectoring exhaust nozzles are performed. Regarding the Reynolds timeaverage equation as the governing equation, the SST k-ω model is adopted to study the pitch and thrust vector angle characteristics and thrust characteristics of the axisymmetric vector nozzle at different geometric deflection angles in the design state of the micro turbine engine. And the simulation results are verified and supplemented based on the vector nozzle characteristics experiments of this type of turbojet engine. The numerical simulation and experimental results show that the nozzle's geometric deflection angle is at between 0° and 20°, the engine thrust loss is small, and there is a good linear relationship among the pitch thrust, the pitch thrust vector angle and the geometric declination.

Abstract:Aimed at the problems that it remains to solve the practical problems in the field maintenance at a certain type of digital electronic engine controller, a comprehensive measurement and control system is designed on the basis of fully studying in the signal characteristics of the DEEC and conducting extensive investigation on the actual needs of the frontline maintenance personnel. The measurement and control system output and acquisition accuracy are ensured to realize the miniaturization, modularization and automation of the system by adopting full digital design, the modifying signal processing circuits of DDS technology, the virtual instrument technology and other means. The test shows that the system is simple and convenient in operation, excellent in loading ability, and high in precision, realizing to take the place of the original measurement and control system. In fact, the measurement precision of the voltage and time can be up to ±0.01 V and ±10 ms, the accuracy of the sensor signal simulation can be also up to ±0.2 Hz, ±0.01 V and ±0.01 Ω. Moreover, the system can also become a good hardware platform for studying the intelligent testing technology.

Abstract:As a special bistatic configuration, the imaging model of spacemissile borne bistatic synthetic aperture radar (SMB-BiSAR) is firstly constructed in vector according to the geometry of satellite and missile the range history and the echo signal model being obtained. Then, the constraint of the missile flight path is analyzed based on the orthogonality of isorange and isoDoppler contours. Finally, the spatial resolution expression of SMB-BiSAR is derived by the gradient of range history and Doppler frequency. The simulations show that SMB-BiSAR has the ability to obtain the highresolution image of target area in the case of large squint angle and forwardlooking by the path planning of missile. In addition, the bistatic geometry has greatly effects in spatial resolution. The greater the bistatic angle between transmitter and receiver is, the worse the range resolution becomes. When both the transmitter and the receiver overlap, i.e. the range resolution, is the best. The larger the angle between moving direction of transmitter and receiver is, the worse the azimuth resolution is obtained. And the azimuth resolution is the best, when the motion direction is consistent.

Abstract:In view of solving the problem of radar targets with rotating structures in ISAR imaging, an ISAR imaging algorithm for targets with rotating parts based on multiple measurement vectors and compressed sensing (CS) is proposed. Through analyzing the Doppler difference between the target body signal and the rotating part signal, a model of the multiple measurement vectors (MMV) to the target body signal in direction of crossrange is established. Since the target body signal has a fixed supportset in direction of crossrange, and the rotating part signal does not have sparsity on this supportset, the body ISAR image of the target can be obtained after the signal reconstruction with the MMV model. On the basis of this, the ISAR image of the rotating part is obtained by inverse Radon transform. The simulation results show that the proposed algorithm is valid.

Abstract:Aimed at the problem that midpoint voltage is out of balance in threelevel active midpoint clamp inverter (ANPC), an improved SVPWM control strategy is proposed. Firstly, the reasons of the point voltage imbalance in the ANPC inverter are analyzed, the midpoint current formed by each vector is analyzed by the space vector synthesis method, and an improved modulation algorithm is proposed by using vector in large vector synthesis. The offset of midpoint voltage in SVPWM modulation mode is analyzed, and a solution is given. On the basis of this, the modified SVPWM control strategy and the traditional SVPWM control strategy are simulated and compared with others, and the advantages of the improved algorithm are analyzed. Finally, the results are verified by the experimental platform. The results show that the modified SVPWM control strategy can effectively balance with the midpoint voltage, and can correct offset to be the midpoint voltage in time, and can make the midpoint voltage weak in fluctuation.

Abstract:In order to solve the problem of multiple access interference of soft spread spectrum signals, this paper adopts the Fast-ICA algorithm to realize the blind separation of soft spread spectrum multiaccess signals of Walsh code by taking advantage of the fact that different users’ information is independent from each other and spread spectrum codes are not related. The algorithm is used to group the received signals according to the known pseudocode rate and pseudocode period, the PCA algorithm is used to realize the dimensionless whitening of the grouped signals to eliminate the correlation between the signals, and finally the Fast-ICA algorithm is used to realize the blind separation of soft spread spectrum multiple access signals and the pseudocode sequence estimation. The algorithm is characterized by blind separation of soft spread spectrum multiple access signals with known pseudocode rate and pseudocode period. The simulation results show that the proposed algorithm can achieve the blind separation of soft spread spectrum multiple access signals and the estimation pseudocode sequences, which is Walsh code of no more than 5 users in the range of a certain signal noise ratio.

Abstract:Aimed at the problem that the communication is comparatively poor in quality under condition of battlefield environments, this paper proposes a clustering algorithm based on the weighted vehiclemounted Ad Hoc networks with the same mobile characteristics. First, the paper calculates the speed similarity and distance similarity of mobile nodes, groups the nodes with the same moving characteristics and close distances under condition of the same cluster, and sets the maximum threshold value of cluster members in each cluster. The paper calculates the speed factor, the distance factor, and the average link maintenance rate. These three factors for each mobile node are weighted in the clustering, and then the largest weight is selected as a primary cluster head, and the second largest weight is used as a secondary cluster. If the primary cluster head once is failure, the secondary cluster head can play a role in the primary cluster head, and the primary cluster head allocates TDMA time slots to each member, improving the communication quality. The simulation results show that compared with the WCA algorithm and the WBACA algorithm, the clustering algorithm greatly improves the stability of the cluster structure and the packet delivery rate.

Abstract:In view of solving the problems of large clock jitter in CDR circuit of SerDes, a new highspeed and high linearity phase interpolator is designed. On the basis of the nonlinear mechanism of the input control code and the output clock phase of the phase interpolator, and through calculating the inverse function relationship between the output clock phase and the weight of the tail current source, array parameters of the tail current source in the phase interpolator are designed accurately, realizing the high linearity relationship of the phase interpolator at high speed and effectively improving the CDR recovery Clock jitter performance. In this paper, the technology of a 22 Gb/s SerDes receiver based on CMOS 65nm process designed is verified. The simulation results show that the linearity of the phase interpolator and the jitter performance of CDR recovery clock are increased by 55.1% and 22.5% respectively.

Abstract:This paper designs a fast locking and low jitter clock and data recovery circuit applied to 28 Gb/s nonreturntozero code highspeed serial optical communication receiver. In order to solve the problem that clock jitter and locking time are difficult to be considered at the same time, the lock detection and discrimination technology is proposed in the circuit with separate proportionalintegral path to realize the adjustable gain of proportional path and enable the loop to lock quickly under condition of low jitter. The simulation is carried out by Cadence Spectre. When the lock detection and discrimination technology is used in the loop, the locking time is about 400 ns and the jitter peaktopeak value is about 2.5 ps. Compared with the other two schemes, the locking time is reduced by 33% and the jitter is reduced by 43%.

Abstract:In this letter, a low RCS patch antenna array based on integrated of radiation and scattering performance has been designed. The 4×4 array is comprised of two different antenna units, which have similar radiation characteristics but 180° reflection phase difference. By using genetic algorithms, aperiodic configuration can be obtained, which makes the scattering field evenly scattered in all directions and effectively reduces the peak value of RCS. Simulation results shows gain of the proposed array is more than 16 dBi. Under xpolarization the monostatic RCS in normal direction is reduced from 8 GHz to 15.9 GHz and the peak reduction is 17.1 dB. Under y polarization the monostatic RCS in normal direction is reduced from 9.5 GHz to 16 GHz and the peak reduction is 13.1 dB. The proposed array was manufactured for test and the measured results show good agreement with the simulation results. The proposed array has characteristics of metamaterial and patch antenna. The RCS is reduced because of phase difference of antenna units without any other metamaterial, which provides a new method to reduce the RCS of the antenna array.

Abstract:Due to the high mobility and multistage mobility of the maneuvering targets, the singlemodel tracking algorithm is difficult to achieve accurate tracking. To solve this problem, AGIMM tracking algorithm based on fuzzy logic and maneuvering detection is proposed. Considering the slow convergence speed of the mesh model of adaptive grid interacting multiple model algorithm and the problem that the maneuvering method is too dependent on the posterior probability of the model，the algorithm introduces fuzzy control to detect the credibility of the maneuver decision. Then, the adjustment method of the model mesh is regiven by the information of maneuver detection credibility target maneuver information and the model posterior probability, which overcomes the shortcoming of the slow convergence speed of the AGIMM algorithm and achieves a fast matching between the model and the target motion mode. Simulation results show that, under different observation conditions, the position tracking error of the proposed algorithm is at least 5.6% lower than that of the AGIMM algorithm, and the position tracking error of the proposed algorithm is at least 15.1% lower than that of the fixedstructure interactive multimodel algorithm. Experimental results show thatthe proposed algorithm guarantees faster model convergence speed and large target tracking accuracy.

Abstract:In view of the requirements of target recognition, the deep learning methods based on neural network are set off. Generally, there is a priori hypothesis of data contained in the indepth learning model, the artificial design aimed at neural network for data needs an abundance of priori knowledge for experts in the field, and has the disadvantages of laborintensive and high time cost. In order to obtain better network performance beyond the personal experience of network design experts, an efficient structure search method, i.e. Differentiable Architecture Search, is adopted. In this method, the search space is broadened to be continuous, and then the performance of the verification set of the architecture is optimized by gradient descent, finding the optimal neural network structure for target recognition. The simulation results show that it is feasible to apply the target recognition method based on neural network structure search to the LSS target recognition.

Abstract:The airborne networks being lack of supporting capability to multitasks, Softwaredefined network brings about a new way to us.In order to reduce the deployment cost of the network, flexible and easytodeploy UAV sare used to form a logically centralized control plane to control the manmachine formation with high dynamic topology changes. In view of the deployment problem of UAVbased controllers, and for the sake of optimizing reliability and deployment cost overhead, a redundant UAV controller deployment strategy based on redundancy deletion is proposed.Firstly, in order to realize the full coverage of the task area to ensure connectivity, the initial deployment is based on the communication range of the UAV.Then, in order to determine and delete the redundant UAV in the initial deployment, according to the deployment constraints and optimization indicators, redundancy decision algorithm based on connection relationship and redundant deletion algorithm based on network connection are proposed.The experimental results show that compared with the global coveragebased controller deployment strategy, the proposed strategy reduces the number of deployed UAVs by 25% while the reliability requirements are satisfied, and the deployment cost of the network is reduced and is more suitable for controller deployment scenario in a dynamic topology environment.

Abstract:Aiming at the characteristics of large-scale combat resource service pools and low service organization efficiency in the command information system under the SOA architecture, a clustering study of service organizations is conducted. Firstly, the attributes of combat resources are analyzed, based on the OWL-S description specification, serviceoriented modeling of combat resources is formed, and an atomiclevel service ontology model of combat resources is formed to realize the cloud service of combat resources. Then a clustering model of combat resource service is constructed, and the simulated annealing method and the K-means method are combined in the genetic algorithm, and a clustering method of combat resource service is proposed. Finally, the simulation is compared with the GS and GK algorithms. The experimental results show that the proposed algorithm can find a clustering scheme with a higher fitness value within an acceptable time, and the standard deviation of the fitness value of the solution results is lower. Compared with GS and GK algorithms, this method has better optimization and stability.

Abstract:Locally repairable codes are a kind of new erasure codes capable of providing information repair ability for distributed storage systems. Aimed at the problem that the researches on the locally repairable codes are still insufficient in ternary field at present, a general method of constructing locally repairable codes based on cyclic codes is presented. Firstly, proceeded from the code length of cyclic codes, the corresponding 3cyclotomic cosets are calculated, and then the defining sets of each cyclic codes are determined through the combinations of cyclotomic cosets, thus determining the distance and locality of codes, and constructing the three kinds locally repairable codes reaching CadambeMazumdar (C-M) bound with 8≤n≤50. Particularly, three kinds of optimal locally repairable codes with small locality are constructed by designing the dual distance through defining sets and by using BCH boundary to filter the circular coset. The relevant construction theory of ternary locally repairable codes is further being perfected by these research results.