This PDF file contains the front matter associated with SPIE Proceedings Volume 10835 including the Title Page, Copyright information, Table of Contents, Introduction, and Conference Committee listing.

A predictive function control method based on the adjustment coefficient is proposed in this paper. For the first order inertia plus pure delay object in industry, the optimal control law of predictive function can be obtained when using a basic function (step function). In this paper, a single adjustment coefficient optimal control law is obtained, in order to improve the control quality, the filter link is added after the control law. Secondly, a single adjustment coefficient and filtering inertial time constant setting method are proposed, that is, the genetic optimization algorithm is used to set the adjustment coefficient, the specific steps of genetic algorithm optimization are given, the performance index design and the range of adjustment coefficient optimization parameters are given. According to this method, the optimal adjustment coefficient can be obtained for any first-order inertial delay object. Finally, the simulation verification of the algorithm is given, and the validity of the algorithm is proved by the experiment.

Recently we have been concerned with locating and tracking targets in aerial videos. Targets in aerial videos usually have weak boundaries due to moving cameras. For the purpose of target detecting, detecting the contour of the target is needed and can help with improving the accuracy of target tracking. Edge detection has assisted in obtaining some advances in this effort. However, noisy images and weak boundary limit the performance of existing contour detecting algorithms. After analyzing the structures and edge maps of a Holistically-nested Edge Detection network, we utilize the highest level side-output and improve the architecture of HED; firstly we cut and resized our images into 400*320 pixels. Secondly, we detected edges using our improved HED network. Finally, the contour of an object is found based on edge detecting in the previous stage. We have significantly decreased time spent by reducing 5 side output layers to only 1 and replacing the fusion layer with a refinement and image processing module which also helps with the result. The experimental results show that our algorithm outperforms the state-of-the-art regarding images with noise and weak boundary.

To solve the biometric technology based on artificial intelligence and deep learning, to improve the system resolution is an important aspect of the development of a complicated sensor. In order to improve the resolution of the imaging system, and achieve the theoretical limit, we introduced the technology principle of super resolution restructure from the point of view on theory and engineering. Several methods to realize high resolution restructure configurations are introduced based on theoretical analysis and engineering practice. Then, three kinds of restructure technologies, that prototype, micro scanning and sub pixel are described, and how to decrease their shortcomings are discussed in detail. The results support theoretical case.

Large scale global optimization problems are closely related to real-life; however, the existing test function sets for large scale optimization problems can not truly reflect the complexity of the actual optimization problem. This paper presents a method for constructing test function sets, it can generate complex test function with different correlation, different deception and different difficulty of solving by adjusting the key parameters such as encoding length, number of groups, equipartition, continuity and the upper and lower limits of the dimensions within the group, it can be controlled by correlation, deception and continuity among dimensions. Using the existing metric correlation index verified the validity of the new construction test functions, and it can effectively simulate the incompletely separable optimization problem with different complexity.

The theoretical framework for neural networks remains incomplete. simulation of the single sample identification process in humans demonstrated that single-sample learning could generate a good generalization model. With the new activation function, we could simulate the frequency response characteristics of different neuronal types and obtained good normalization. Our study demonstrated the similarities between artificial neural networks and human brain processes.

With the successful introduction and popularization of Kinect, it has been widely applied in intelligent surveillance, human-machine interaction and human action recognition and so on. This paper presents a human action recognition based on multimodal information using the Kinect sensor. Firstly, the HOG feature based on RGB modal information, the space-time interest points feature based on depth modal information, and the human body joints relative position feature based on skeleton modal information are extracted respectively for expressing human action. Then, the three kinds of nearest neighbor classifiers with different distance measurement formulas are used to predict the class label for a test sample which is respectively expressed by three different modal features. The experimental results show that the proposed method is simple, fast and efficient compared with other action recognition algorithms on public datasets.

Deep learning has strong abilities in finding and expressing characteristics of pictures. Recent years, with the arrival of big data era and the development of computers, deep learning has made great breakthroughs and become the focus of the field of computer vision. First the history and classification of deep learning are presented. This thesis also introduces the basic theory of typical deep learning models on computer vision, which include convolutional neural network, recurrent neural network and generative adversarial network. And then summarizing the research situations and progress of deep learning on image classification, image detection, image segmentation as well as video recognition and prediction. Finally, the development and trend of deep learning in the field of computer vision are analyzed. The combination of convolutional neural network and recurrent neural network will be a good choice for video recognition and prediction, which still has a big gap between human beings cognition. And it is the generative adversarial network which has strong ability to generate new samples based on the potential distribution will play an important role in computer vision.

For fault diagnosis of automotive assembly, the experience knowledge of experts is still the primary means which is inconvenient and time-consuming. This paper proposes a novel fault diagnosis method for automotive assembly based on optical coordinate data and machine learning. First, after obtaining large amounts of measured data from sensors, a feature selection method which is based on the subset-level score is performed. Because high-dimensional data will result in high computational cost and irrelevant and redundant features may also degrade the performance of fault diagnosis. The feature selection method which can be classified into filter methods needs to choose a certain evaluation criterion first. By using a fast iterative algorithm, we can finally find the optimal subset of features. Second, for the nonlinear relationship between measured data, we introduced well-known kernel methods to efficiently improve feature extraction effect. In machine learning, kernel methods are a class of algorithms for pattern analysis, whose best known is the support vector machines. Last, Bayesian inference classifier is built to recognize the fault pattern. The experimental on real production process data of automotive assembly show that the proposed method can greatly enhance the diagnosis accuracy.

Violence detection in videos is a challenging task which has gotten much attention in the research community. In this paper, we propose a three-stream network framework for violence detection in binocular stereo vision. To capture the complementary information from the video we adopt the appearance, motion and depth information. The spatial part, we use the RGB as the individual frame appearance. Then, we use the sparse stereo matching method to extract the feature points and obtain the vision disparity of the point. The 3D coordinates of the points are calculated through the standard 3D measurement theory. The 3D motion vector conveys the movement of the camera and the objects as the motion information. Besides, the depth information flow is the third input of the network which can improved recognition rate.

Human hallmarks, as an important biometric traits, contain discriminative characteristics information to support person identification due to its saliency in visual attention. A sketch of a tattoo can be drawn based on the description provided by an eyewitness or the victim. In this paper, the authors propose a novel hallmark photo-sketch recognition method using a bag-of-words (BoW) methodology for retrieving the photos in the database based on a query sketch or image drawn by an artist. And this article discusses the number of visual words generated which is based on the number of clusters, including vocabulary size, weighting scheme. Extensive experiments are conducted on a sketch database including 100 tattoo sketches drawn by three different subjects. Each sketch has a corresponding tattoo image, which is downloaded from the internet. It is an effective and viable application to develop operationally-relevant automated image-based tattoo recognition applications for pertinent law enforcement operational scenarios.

During the "13th Five-Year", with the implementation of major special tasks such as the space station and the three phase of navigation, the development of the spacecraft has gradually taken on the characteristics of large quantity, great variety, short cycles and heavy tasks. The traditional method of using theodolite for collimation measurement has long time of measurement, low level of automation and high occupancy rate, which is unable to meet the development requirements of high reliability and high efficiency of the current spacecraft. According to the requirement of equipment assembly accuracy measurement and the characteristics of field implementation in the process of spacecraft assembly integration and test (AIT), a method for carrying theodolite by robot to measure the assembly accuracy of spacecraft equipment is proposed, which takes full advantages of flexible and high automation level of industrial robot. After experiments and field application verification, this method can greatly improve the measurement efficiency and automation level of spacecraft assembly on the basis of ensuring high-accuracy measurement, the corresponding system has been successfully applied to the development process of spacecraft such as China’s space station, BeiDou navigation, remote sensing and so on.

As the requirement of satellite antenna’s planar near-field test, the scanning plane, the air-bearing based platform and the satellite antenna should be measured and adjusted with high precision. In order to improve the measurement accuracy and efficiency for large scale objects, a new measurement method is proposed based on laser radar. Target ball measurement, scanning measurement and single point measurement are respectively used to measure the scanning plane, the air-bearing based platform and the satellite antenna. The working principle, automatic measurement method and measurement error are described in detail. The experimental results show that the measurement error is less than 0.1mm for 3D point coordinate and 10” for attitude angle. Compared with traditional theodolite measurement method, this method can realize automatic measurement with higher precision and efficiency.

In this paper, the uncertainty model of laser radar and Metrascan detection is established, combined with the the accessibility of laser radar in the actual measurement, the accuracy of the measured network is simulated and analyzed, and a digital measurement method, which combines the advantages of two advanced measurement techniques, Metrascan and laser radar, is proposed. This method of netted measurement for different digital equipment, uses a high precision laser radar coordinate system as the reference transformation coordinate system, improves the measurement precision of large-size measurement of Metrascan, thus, which realizes the high precision and high efficiency measurement for the digitization of the aircraft shape.

Laser trigonometric displacement sensor has characteristics of high efficiency, non-contact and large-scale measurement range, when coupled with scanning system, it can be widely used in profile measurement of complex workpiece surface. But for large steep workpieces, the axis of incident light emitted from the sensor can’t be perpendicular to its surface, accuracy will be largely degraded by this inclination angle. Also, the relationship between error and influencing factors dominated by inclination angle is a nonlinear function. If the influence of measuring distances is taken into account, the relationship becomes a multivariate mapping. So an improved multi-layer BP neural network is proposed to compensate for errors. This paper uses the genetic algorithm to optimize the initialization parameters of the network, while using adaptive training method to optimize the convergence process and adjusting the learning rate, increasing the momentum item to avoid falling into local extreme points. Besides, the laser displacement sensor of Keyence LK-H020 is used to obtain the measurement data and the error was obtained by comparing with a grating ruler with a precision of 10 nm. Based on the simulation and experimental results, the method can reduce the error from 3.8 μm to 0.5 μm when inclination range is from 0° to 8°, and from 7μm to 3 μm when the angle is from 0° to 50°. The results prove effectiveness, generalization and robustness of the algorithm.

The research focuses on measuring the influence of V-defect on wedge waves propagating along line wedge tip by using laser ultrasound technique. Generally, wedge has more or less defect or damage on the tip, which may result in break and bring economic losses. Thus it is necessary to investigate characteristic of wedge waves propagating along line wedge with defect. The wedge waveguide models with different defect depth were built by using finite element method. Multiple mode wedge waves were observed through B-scan. The open of defect is 0.1mm, and the depth is 0.01mm, 0.05mm, 0.1mm, 0.2mm, and 0.3mm, respectively. It was seen that both reflected and transmitted waves were observed. Due to the dispersion characteristics, we observed the reflected and transmitted A₁ mode separated from A₂ mode, which can be used to determine the width of V-defect. Meanwhile, models of V-defect with different depth are also built. We had found that wedge waves are totally reflected and there is no transmitted wave observed as the depth is bigger than 0.3mm.

Absolute measurement with Phase Measuring Deflectometry (PMD) is gaining importance in industry, but the accuracy of deflectometry metrology is strongly influenced by the level of calibration. In order to improve the accuracy of the PMD to a level where it competes with interferometry, a reference calibration process is commonly carried out to carefully calibrate the systematic errors. The systematic errors obtained by measuring a high quality reference surface can be subtracted from the measurement of a test surface to get its actual surface, however, it could introduce the surface error of reference into the measurement. To alleviate this problem, this paper introduces a technique named “rotational shear phase measuring deflectometry”, this technique have the ability of removing the rotationally asymmetric systematic errors from the test surface without using a reference surface. The validity of this technique has been demonstrated by simulation and our experimental results.

This study presents an automated system to effectively identify different noise types and levels of interference fringe pattern images. The key idea involves feature extraction from noise samples using extra filters and multilayer neural network. Besides median, wiener and homomorphic filter, NL-means filter is used to separate noise samples based on non-local self-similarity of fringe patterns. Statistical methods like kurtosis and skewness are extracted and used for neural network learning. The system is capable of accurately classifying the type and level of noise of fringe patterns and specific filter can be applied. The experiment result shows that the accuracy of high noise level is still over 82%. We introduce non-local fractional-order diffusion equation filtering method for high level Gaussian noise corrupted electronic speckle pattern interferometry fringes denoising. The proposed method is based on partial differential equation (PDE) and non-local methods. The first term of the energy functional is nonlinear P-M function which can remove noise meanwhile preserve edges. The second term is fractional order total variation energy functional, it can use the selfsimilarity of fringes pattern and improve the quality of the denoised image.

Point to the low precision issue measuring remote target with vision system directly, the characteristics were analyzed. Firstly, the system platform measuring remote target with unmanned aerial vehicle was built. Secondly, the vision system was calibrated with high precision. Finally, the measuring precision of vision measurement was tested by the experiment for the remote target. Then the feasibility measuring remote target with high precision by this method was evaluated and the some measurements were proposed to improve the brought problems in this system.

In the ultrasonic testing of submarine pipelines by using guided waves, wave energy leakage is a main reason of signal decay. For overcoming the decrease of energy attenuation, the propagation of guided waves of immersion plates is studied in this paper. The dispersion equations of guided waves is numerically solved. Then the appropriated modes of which phase velocity are small or large are selected for optical Schlieren visualization and propagation of leaky waves is discussed. It is shown that selecting some modes of which imaginary part are small can retard guided wave decay and extend length of testing.

Bars or tubes are widely used in all kinds of equipment such as automobile shafts or power plant boilers. Effective detection of internal transverse defects is very important to the security of the equipment. In this paper, the time-delay algorithm for the detection by using the combination of a phased array ultrasonic probe and a wedge was discussed. Firstly, a coordinate system was established and the position of each element was calculated. The focal point position of phased array ultrasonic wave was specified. Secondly, based on the geometric acoustics assumption, the equation of the sound wave incident location of each element and its scope on the interface between wedges and artifacts were deduced by using Fermat's principle. The numerical method was implied to solve the equations. And the delay time of each element of the array was computed. Finally, the computed delay time of each element was imported to the finite element model and the acoustic beam was simulated. The result shows that the detecting waves can be focused to the specified position by using the calculated delay time. This paper provides a method to calculate the delay time of each element of phased array probes for detecting the transverse defects of cylindrical surface artefacts.

The durability of photovoltaic hollow glass modules for buildings is an important index. In accordance with the Chinese national standard of the hollow glass (GB/T11944-2012) and referring to the international standard (IEC61215:2005) on test method for PV module in simulated environments, we have designed and conducted an accelerated durability test of photovoltaic hollow glass modules. Based on analysis of the test results, the paper puts forward several critical technical points that should be noticed in the testing process, and points out the failure modes and its mechanism, and would thus help to improve the reliability and durability of photovoltaic hollow modules. Keyword: photovoltaic hollow modules, high temperature and humidity, durability, molecular sieve.

An interest has been attracted in unmanned aerial vehicle (UAV) for its widely practical applications, which can also deal with some difficulties in traditional electromagnetic compatibility (EMC) testing. The UAV is regarded as a flying platform carrying a generator source or a receiving monitor to perform the measurement of electromagnetic susceptibility or electromagnetic radiation emission. Several typical testing scenarios are presented in this paper, and the methods of their realization are analyzed in detail. Although there exists some challenges in the flying test due to the characteristics of the UAV, some effective measures can also be taken to improve the measurement.

In view of the complexity and variability of bathymetric data, the paper introduces a new algorithm named DAE-WGAN to construct sea bottom trend surface. This new model is an alternative to traditional GAN training method, combined with the advantages of Denoising Autoencoder (DAE) and Wasserstein Generative Adversarial Network (WGAN). Firstly, the network structure is introduced in detail, in which the critic (or ‘discriminator’) estimates the Wasserstein-1 distance between the generated-sample distributions and the real-sample distributions, and optimizes the generator to approximate the minimum Wasserstein-1 distance, which effectively improves the stability of the adversarial training. Moreover, the generalized Denoising Autoencoder algorithm is added to train the back-propagation process, having a better ability of dimensionality reduction, which improves the robustness of the whole algorithm. Then, using two different types of bathymetric data (seabed tiny-terrain data and Electronic Nautical Chart data), we had long-time experiments to train the DAE-WGAN till optimality, and got the better sea bottom trend surface. Finally, by comparison with other GAN models (such as InFoGAN, LSGAN), the results show that the proposed method has an obvious improvement in accuracy, stability and robustness, and further illustrate the feasibility of this method in bathymetric precise data processing area.

Sensor task allocation plays a great role in military, environmental science, medical health, transportation and other fields. In order to make rational use of limited sensor resources, a multi-sensor multi-target task allocation method based on an improved firefly algorithm (FA) is proposed. In the algorithm, the initial position of firefly individual in firefly algorithm is optimized to speed up the search optimization procedure. In the process of constructing efficiency function, position constraints, sensor monitoring ability constraints and target threat degree constraints are considered comprehensively, leading to a more realistic multi-sensor multi-target task allocation algorithm. The analytic hierarchy process (AHP) is used to construct the target threat measure. The simulation results show that the proposed algorithm is more efficient than the standard particle swarm optimization algorithm (PSO) and the standard FA, that is, the sensor task allocation is more reasonable, and the task allocation time cost is also shorter than the other two algorithms.

The rapid development of drones not only brings convenience to life, but also triggers various types of security issues. For this reason, in recent years, China's major industrial sites and transportation hubs have been actively involved in the construction of security systems for drones to ensure safe operation. Photoelectric alarm system is one of the most important components of the security system. Its purpose is to identify the target and obtain legal evidence. This article describes an unattended optoelectronic warning system for drones that can simultaneously output visible and infrared image information and automatically identify the target type. This system has been applied in important industrial sites in China and has achieved phased results. Its successful operation lays the foundation for the development of the subsequent intelligent photoelectric warning platform.

Free space propagation loss of aeronautical radio navigation station is calculated according to transmitting antenna height, receiving antenna height, frequency, distance between transmitting antenna and receiving antenna, comparing with traditional free space propagation loss algorithm. The data calculated from the algorithm is checked by flight check data combined with site data around navigation station.

Convolutional neural network(CNN) has achieved great success in optical image recogniton, and has been applied in research of synthetic aperture radar(SAR) automatic target recognition(ATR) recently. However, in real SAR systems, SAR images are usually compressed for transmission due to the limited wireless bandwidth, and few researches have evaluated the influence of image compression on SAR ATR task. In this paper, an efficient CNN architecture based on inception module and batch normalization layer is proposed for SAR ATR tasks, and the impact of image compression on SAR ATR is evaluated based on the proposed model. The experiments are based on MSTAR dataset, and the test images are compressed by Set Partitioning in Hierarchical Trees(SPIHT) algorithm with different compression ratio. Experimental results show that the proposed CNN model achieves a state-of-the-art classification accuracy of 99.29% on original MSTAR dataset, and can still get high classification accuracy above 90% even SAR images in test set are compressed by nearly one hundred times, which reveals that moderate compression of SAR images has little influence on SAR ATR tasks, and this validates the practicability of applying the proposed model to real SAR ATR systems.

Accuracy is highly important on autonomous robots. In this work, we propose a novel visual-inertial SLAM with stereo camera and IMU, which construct sparse map and estimate the camera poses accurately. The camera and IMU data are tightly coupled by nonlinear optimization. pre-integration is used to integrate rotation, velocity, and the pose matrix. A serious techniques are adapted to feature extraction, keyframe selection select keyframes, and loop closure. In addition, the system can run real-time on standard computer. The system localization accuracy can arrive centimetre-level especially in a large scale environment, and system is robust. We elevate the system on public datasets to compare other visual-inertial SLAM approaches; our system achieves better accuracy and robustness.

With the development of science and technology, unmanned aerial vehicles(UAVs) are widely used in various industries, and the research for the Minitype UAVs have been widely studied. This paper introduces a design scheme that the brain controls UAVs system based on cloud platform. Brainwaves of the subject's occipital lobe were collected by visual evoked, which transmits to the mobile phone via Bluetooth acquisition, data processing, and a fixed encoding format is used to control the flight of the aircraft, Environmental monitoring platform is loaded simultaneously, Users can remotely observe the flight status, environmental conditions and so on by the browser. The scheme has a good prospect of application.

An appropriate impact angle constraint can not only decrease the approaching velocity but also enlarge the effective collision area of hypersonic interception, improving the interception probability. Based on the principle of quasi parallel approaching, the impact angle constraint is converted into line-of-sight angle constraint according to the initial estimation of target velocity and velocity vector. The guidance law with line-of-sight angle constraint is presented via variable structure control. On the basis of qualitative analysis on interception performances, the fuzzy strategy is proposed to generate the appropriate impact angle and weight coefficient of constraint. It is concluded that, the proposed guidance achieves to guide the line-of-sight to close to zero as well as reach the expected impact angle that automatically generated by the fuzzy controller. Moreover, the miss distance and applicability to target information estimation error are satisfied.

In order to meet the increasing workload demands of geographical conditions monitoring, solve the problems of traditional monitoring equipment such as unable to intelligent connection each other，single monitoring, complex monitoring environment, various types of monitoring objects, unreasonable allocation of monitoring tasks, the higher and higher processing efficiency and precision monitoring data, this paper proposes an integrated intelligent monitoring platform for the operation of the dispatching service center with the cloud server as the brain, and land, sea, aviation and thematic as the service limbs. It uses the Internet of things technology, wireless sensor networks technology, and radio frequency technology to realize the cooperation between the devices to complete automatic data collection and automatic remote transmission. The service center, based on cloud computing, makes intelligent control, decisions, and completes tasks such as planning, allocation, data processing, data publishing, and platform authorization. In this paper, some achievements of single monitoring method are presented, and some technical principles are introduced. Some Suggestions will be provided for future intelligent monitoring by the research of this platform.

The traditional measuring method is incapable of meeting requirements of increasing measuring work due to the development of artificial intelligence and big data technologies. A novel development concept of the intelligent measuring was proposed in this paper to conquer drawbacks, such as lacking of intelligent relevance, possessing the only one kind of measuring method, low efficiency and accuracy in measuring data processing, of the traditional measuring equipment facing the challenges of complex environment and diverse objects of measurement. Intelligent unmanned system composed of technologies of internet of things, wireless sensor network and radio frequency are applied to aviation, land, ocean and other occasions, which makes automatic data collection and long-distance transmission done by remote equipment’s teamwork come true. Based on the cloud computing technology, tasks receiving, data processing and publishing, and data authorizing were conducted by the service center. The present work demonstrates some results obtained by the only one kind of the current measuring method. Further, theory of some technologies and methods was introduced and importance of intelligent unmanned system on the development of intelligent measurement. The current research work provides some suggestions of help on development of intelligent unmanned equipment and progress of measuring technologies.

Simultaneous localization and mapping (SLAM) is a hot issue in the field of unmanned system. It is an essential task for the development of autonomous robots.The research achievements and historical development of SLAM are reviewed in this paper. And the key technologies such as detection and matching of feature points, selection and matching of keyframes, loop closure detection are summarized. Furthermore, the advantages and disadvantages of current SLAM methods are discussed. Finally, the research hotspots and trends of advanced SLAM are summarized and prospected.

According to the characteristics of the random errors of MIMU, an Allan variance analysis method fused with genetic algorithm is proposed, which can effectively evaluate the different random errors. Firstly, how to analyze and identify the errors of inertial devices by Allan variance analysis method is introduced in detail. Then, according to the characteristics of genetic algorithm that can achieve global optimum, an Allan variance analysis method fused with genetic algorithm is proposed. Finally, by long-time experiment to test the MEMS inertial devices of three different manufacturers, the measured data of gyro and accelerometer are processed and compared respectively, and the numerical results of each random error have been calculated, which proved the validity of the method. This method combines genetic algorithm with Allan variance analysis method, providing a new idea for the theoretical study of random error field in MEMS inertial devices.

Counteracting LSS-Target (the Low altitude, Slow speed Small Target) has become a hot topic in security field in recent years. However, some technical means are not fully mature. A kind of fully autonomous and agile response anti-LSS-Target system has been proposed. Through one approach based on deep learning, a convolution neural network (CNN) is constructed and trained to realize the effective recognition of UAV. The tracking model of UAV is built based on discrete Kalman filter algorithm to achieve long-term tracking in the field of view. The test results show that after identifying the target UAV automatically, the system locks the target and tracks it steadily.

In synthetic aperture radar (SAR), many of the biased baseband Doppler centroid estimates are caused by partial exposures of bright targets. To avoid this problem, this paper proposes a new estimation algorithm to remove the effects of partial exposures under non-homogeneous environment. This algorithm is implemented by three steps. Firstly, concentrate the energy of each target into a few pixels through azimuthal compression using frequency domain matched filter. Then, extract the homogeneous environment through the spectral distortion and contrast, and eliminate the strong interference target. Finally, estimate the centroid of the averaged extracted power spectrum using a circular convolution method. Theoretical analysis and experimental results show that the proposed method can effectively eliminate the interference targets and accurately find the Doppler centroid under non-homogeneous environment with strong interference. Moreover, the operations of the proposed algorithm need no iterations, thus the computational load is greatly reduced compared with the traditional iterative methods.

The phase gradient autofocus (PGA) technique is applied over a wide range of imagery and phase error corrections for synthetic aperture radar (SAR). In this paper, we propose an improved PGA method to increase the phase error estimation accuracy by selectively increasing the pool of quality synchronization sources. This improved method operates mainly in three steps. Firstly, after compressing the deramped data by taking FFT, strong targets can be identified and selected out according to the pixel magnitudes over the two-dimensional (2-D) image. Secondly, sort the selected targets by computing their contrast in the azimuth direction and select out the ones with good contrast. Thirdly, sort the selected targets by computing their peak signal-to-noise (PSNR) in the azimuth direction and select out the good quality targets. After these three times filtering, those selected scatterers are optimal in terms of having sufficient signalto-noise energy with negligible impulse response interference to the neighboring targets and clutter scatterers. With these selected quality scatterers, the proposed modified PGA method can achieve a better focusing performance. Advantages and effectiveness of the proposed algorithm are verified on the real stripmap SAR data.

As a kind of game-changing combat style, the operation of the UAV swarm in cyberspace is characterized by high effective cost and high concealment, which make it play an important role in the future competition of cyberspace. Due to the lack of actual combat experience and high confidentiality of the cyberspace operation and the UAV swarm operation, the modeling and simulation (M and S) of the operation of the UAV swarm in cyberspace is the vital method for planning the UAV swarm system architecture and designing the tactical methods of the UAV swarm. At present, the research on the M and S of the operation of the UAV swarm in cyberspace is just starting from home and abroad, and there is no corresponding modeling method and concrete model. Taken the Integrated Air Defense System (IADS) as the operational object, the guiding ideology of the M and S of the operation of the UAV swarm in cyberspace was researched from the height of the joint operations, which includes the characteristics of the individual autonomy, the group cooperation, the offensive and defensive antagonism and the electromagnetic-cyber the operation. Then, the model characteristic and elements was analyzed, such as the model system architecture, the operational rule and the intelligent algorithm. On this basis, the "action model pool", "operational rule pool" and "intelligent algorithm pool" would be constructed. Afterwards, the modeling method of the operation of the UAV swarm in cyberspace which based the “Rule+ Respond+ Algorithm” (RRA) is proposed, including the modeling framework and main process method of the RRA. The main process included the military concept modeling stage, the operational rules refining stage, the operational rules judgment stage, the action model response stage and the intelligent algorithm solution stage, that the five stages were in sequence, also the main work in each stage was analyzed. At the level of system of system M and S, the modeling focused of the abstraction, extraction and description of the operational rules of the UAV swarm in cyberspace. The description and construction methods of the operational rule are researched, such as the BNF format, the process format, and the operational rules tree graph. Finally, based on the future joint operations, an example of the operation of the UAV swarm in cyberspace is given, then the operational effectiveness is analyzed. The example showed that using nine UAV Swarms to attack IADS in cyberspace would reduce the normal power range of the sensor network by 90% and the warning time by 80.3%.

For the high cost of practicality experiments and the low precision of available model data cause the deficiently of the effectiveness of the aeromagnetic detection result, a magnetic field simulation method of submarine aircraft model is proposed based on COMSOL Multiphysics finite element software. First, the submarine is equivalent to the hollow cylinder model, and the magnetic field distribution of the distance of 5 times the length of the submarine is received through the simulation. Then the aircraft geometrical model is established to analyze the mechanism of aircraft jamming magnetic field, and the four-way flight scheme of compensation flight state is proposed to get magnetic field simulation data. For the simulation difficulties of large-scale model in the detection flight state, a simulation method of model separation and data superposition is proposed to get the component and the total data of magnetic field. Finally, the aircraft interference magnetic field modeling and the use of ridge estimation method to compensate for compensation flight and detection flight magnetic field data. Four-course flight data compensation accuracy within 0.5nT and detection flight data after compensation can be distinguished the submarine signal. The correctness of the finite element simulation method and the magnetic field data is verified, which is of great significance to the study of aeromagnetic deection simulation.

To improve the survivability of multiple ground vehicles when they are maneuvering under the reconnaissance of satellites, a survivability planning modeling method and a cooperative task area allocation method based on genetic algorithm were proposed. Firstly, the problem of task area allocation for multiple ground vehicles is depicted and a framework to solve the problem is established. Then satellite reconnaissance model and cooperative task area allocation model are established. And a customized genetic algorithm is adopted to perform cooperative task area allocation to minimize the total mission time and the exposure time under the reconnaissance of satellites. In addition, a survivability assessment framework is established to evaluate the survival status of ground vehicles. Finally, simulation results show that the proposed method can improve the overall survivability of multiple ground vehicles effectively.

Mission resources scheduling and network topology of UAV multi-platform avionics system represents a feature of time and space isolation, and traditional aeronautical Ad Hoc network research and mission resources scheduling research is relatively independent. Aeronautical Ad Hoc network research lack of correlation of missions and resources, which makes the networks of UAV multi-platform avionics system cannot satisfy the dynamic mission demand to resources. In order to solve this problem, we propose a mission driven network and resource mapping architecture of UAV multi-platform avionics system based on a network structure from the perspective of network topology. It studies the quantitative mission, mission scheduling, node mobile model and network dynamic clustering method, through these studies the mission requirements and platform resources of UAV multi-platform avionics system are correlated. And the mission oriented adaptive networks architecture of UAV multi-platform avionics system is proposed, it provide a reference for the mission resource management and the aeronautical Ad Hoc network study of UAV multi-platform avionics system. The proposed network architecture could adaptive adjust the network clustering structure according the mission requirements of UAV multi-platform avionics system, and support the resource connecting and sharing of UAV multi-platform avionics system.

Wireless Sensor Network (WSN) technology has been applied more and more widely and node localization is an important aspect of it. Bounding Box localization algorithm has been used in many cases. The purpose of this paper is to study the localization accuracy of the three different strategies based on Bounding Box localization algorithm and to explore the influence of the two parameters, the number of the anchor nodes and communication radius, on the localization accuracy. Firstly, the paper illustrates the principals of three strategies according to whether the unknown nodes that have been located will participate in locating other unknown nodes or not. Then the simulation condition is set, and the average localization error is gotten using three strategies respectively. The result shows the localization accuracy of Strategy C is the highest. Finally, the paper studies the influence of the number of the anchor nodes and the length of the communication radius on the localization accuracy and gives the optimal number of anchor nodes and communication radius when used in practice.

The integrated Reconnaissance/Strike UAV has very important significance for mastering the battlefield initiative and significantly improving the performance of future commanding combat systems. In this paper, staring with the requirements of military actual combat applications for the discovery and identification of battlefield military targets, the current status and development trend of intelligent target recognition technology for the integrated reconnaissance/strike UAV are summarized, as well as the key technical issues are analyzed and preliminary test results are given.

The signal simulator is an important semi-physical simulation equipment, which is often used to test the performance of the Synthetic Aperture Radar (SAR) fully in a lab environment. Two issues need to be focused on, one is the synchronization between the SAR and the simulator, the other is the application range of the simulator. In this paper, the impact of the synchronization errors on the test is analyzed, the corresponding mathematical models are derived, and the indicator of phase synchronization is presented. Then a scheme of high-accuracy SAR signal simulator based on adaptive synchronization technology is proposed. By recognizing and processing the input frequency from different SAR systems based on coherent methods, the frequency signal can be converted into an important reference clock that the simulator can adopt directly, so the simulator can be adaptive to different frequencies covering almost all kinds of SAR systems. Furthermore, by means of analyzing the difference between the input frequency and the inner reference clock of the simulator, digital phase processing unit integrated in the simulator can obtain the phase synchronization of the SAR system. The experimental results show that the method is reasonable and effective in the end.

For the purpose of improving the range resolution of SAR system under the existing technical conditions, multi-channel synthesis is one of the main technical approaches. However, this method can bring the problems of amplitude-phase distortion and channel synchronization among different channels. In this paper the structure of multi-channel signal synthesis is built, the error characteristics of synthetic wideband signal and the errors’ influence on the synthesis are analyzed, and a mathematical model of channel errors is established. Then an adaptive channel error correction method, which is based on the improved internal calibration method, is proposed. By adjusting the input power of the exciting signal, the low-power transmitting signal is made to work in a stable saturation area, where the amplitude-frequency and phase-frequency characteristics of the signal are constant. By means of pre-distortion technology, the transmission characteristics of the channels can be obtained, and the phase-mplitude error in the multi-channel can be eliminated and corrected.The experimental results show that the method is reasonable and effective in the end.

Doppler Beam Sharpening (DBS) technology and Short time FFT Transform (STFT) are both widely used in unmanned airborne scanning radar for Ground Moving Target Indicator (GMTI) mode, by which radar acquires medium-resolution images of vast areas and detects moving targets in the area. Traditional DBS method is independent of GMTI, hence data must be copied and processed individually, which consume more memories and computation time. We propose a new method to synthetize a long aperture FFT result for DBS imaging with several short aperture FFTs generated by STFT, which is suitable for real-time DBS processing. For data after range compression and STFT in the azimuth direction, a proper up-sampling in the Doppler domain is performed firstly. Secondly, up-sampled data of each sub-aperture are multiplied a function with linear phase to shift corresponding signal of time domain. Finally, we sum data over sub-apertures and get a long aperture FFT result, which equals to the direct FFT result in traditional DBS processing. The effectiveness of the proposed algorithm is validated by simulated data.

Imaging and parameter estimation of moving targets in airborne single-antenna high-resolution synthetic aperture radar (SAR) system is affected by motion errors, which cause defocusing and dislocation of moving targets. Motion errors derive from platform flight deviation and unknown target velocity and estimating all motion errors once is difficult because both aspects influence each other, especially in unmanned aerial vehicle (UAV) systems. We find that the platform flight deviation has the same effect on stationary targets and moving ones. Exploiting this similarity, we propose a novel ground moving target motion compensation method. Platform flight errors are extracted from stationary targets by autofocus algorithms and compensated for moving targets. And then Hough transform (HT) and Map-Drift (MD) technique is used to estimate the radial velocity and the along-track velocity, and range cell migration correction (RCMC) is completed by estimated velocities. Furthermore, PGA technique is adopted to estimate and correct residual phase errors. The effectiveness of the proposed algorithm is validated by the real data.

In the traditional systematic calibration of inertial devices, the calibration path is designed by the reclosing of the IMU sensitive axis and the turntable rotation axis. The rotation axis is 1 times per rotation, and only 2 sensitive axis positions are changed. In order to effectively motivate the error items of inertial devices, an IMU off-axis installation structure is designed. The rotation axis of the turntable is 1 times per rotation, and 3 sensitive axis positions can be changed at the same time. On this basis, a systematic calibration scheme for inertial devices based on IMU off-axis transposition is designed. A 30 dimensional system errors model including 24 error terms, such as gyro and accelerometer constant errors, installation errors and scale factor errors, is set up. The observability of each state in each calibration path is analyzed by piece-wise constant system (PWCS) and singular value decomposition (SVD) method. Compared with the traditional classical systematic calibration scheme, the proposed scheme can not only realize the full dimensional observation of the state, but also the observability of each state is higher on the whole than that of the traditional scheme. In the process of using filter to calibrate inertial devices, the calibration time is shorter and the precision is higher, and the number of turntable axes required for the proposed scheme is changed from three axis to double axis, which reduces the requirements for hardware conditions.

Error calibration is an effective way to improve the measuring accuracy of MEMS Gyroscope, and the scale factor is the key parameter in it. In order to test the scale factor of various MEMS Gyroscopes and make effective evaluation, the thesis put forwards a fast test method of scale factor on the basis of National Metrology Technical Specification JJF 1535-2015 (MEMS Gyroscope Calibration Specification). First of all, it analyzes the cause of scale factor error and establish the corresponding error calibration model. Secondly, it designs the speed calibration experiment in the range of [-160 degree /s, 160 degree /s] angular rate. In view of the miscellaneous data acquisitive operation in the National Metrology Technical Specification, a rapid test method of scale factor is proposed by reasonable reduction of rate calibration points. In the end, it selects three different types of STIM 300, MTI100 and 3DM-GX4-25 to do experiment. The results show that the method proposed in this thesis is basically consistent with the results of National Metrology Technical Specification, and the workload of data acquisition is reduced by half, thus the rapid test of scale factors is realized.

In future battlefields, UAVs play an increasingly important role and need to perform a variety of tasks. The UAV avionics system is also becoming more and more complex. In order to achieve flexible deployment and utilization of UAV avionics resources for various tasks, the composition and workflow of the UAV avionics system are firstly studied. Then the servitization process is studied based on UAV avionics resources. At last, extraction methods from resource to service and service management mode are described, to provide support for the UAV to realize the intelligent and autonomous operation.

This paper presents the design, data processing and experimental results of the C band SAR developed at the Institute of Electronics, Chinese Academy of Sciences (IECAS). This C band SAR aims at oil spill monitoring and adopts a compact, low power consumption and lightweight design based on microstrip antenna, solid state power amplifier and 2- axis gimbals allowing installation on UAV. To further process SAR images, the ground support system is employed to accomplish SAR image processing, geometric correction and geocoding, annotation and shape characteristics analysis of pollution. The first flight test has been performed using C band SAR on Harbin Y-12. Large area SAR image with pollution region of sea surface was obtained. With the ground support system, perimeter, area, shape complexity factor and center latitude and longitude coordinates of pollution region are presented in the monitoring report. The experimental results demonstrate the feasibility of C band SAR and the usability of ground support system in marine monitoring. In the future, the C band SAR should be mounted on the UAV platform to perform oil spill monitoring test.

In addition to the inherent speckle noise, the Synthetic Aperture Radar (SAR) images exhibit low contrast and low signal-to-noise ratio (SNR), in some cases such as ocean monitoring with fierce wind. Given this problem, a novel method for target feature enhancement based on Discrete Shearlet Transform (DST) and multi-scale analysis theory is proposed in this paper. This approach captures the intrinsic geometrical features of target with discontinuities points in the SAR images effectively. In this work, the SAR image is decompose in multiple scales to get different sub-bands, the shearlet coefficients of images in different sub-bands with different directions are fusion. As the scale increases, the shearlet coefficient maximum of the target also increases, while the shearlet coefficient maximum of the speckle and clutter decreases. Therefore, the high-frequency features of different scales in different directions are fused, which makes not only the target enhanced but also the speckle and clutter suppressed. Experiments on ocean SAR images with strong speckle and clutter have been performed. Comparison with traditional wavelet approach, the results demonstrate that the proposed method is competitive in target feature enhancement and clutter suppression.

This paper introduces the research situation of the unmanned cluster’s battle on the sea, analyzes the basic composition and architecture of the unmanned cluster battle system. It gives the operation process of the unmanned cluster. Fromwhich, the key technologies involved in developing the command and control of the unmanned cluster was put forward.

Based on the operational requirements of the future of naval warfare and electronic warfare, according to the composition of the electronic systems of maritime battle groups, analysis of the overall capacity needs of the electronic information warfare. The way of electronic warfare UAV is proposed for in information confrontation. And we analysis early warning radar’s simulation model and electronic warfare process. Then we analysis the detection model of early warning radar in noise jamming, in densely false targets jamming, in repeater jamming and Integrated suppression of jamming. This paper build a simulation platform based on the demand for electronic countermeasures system. Simulation experiments, by comparing the assessment indicators, the results illustrate some typical problems of the penetration of electronic warfare UAV to aircraft carrier battle group, penetration aircraft carrier battle group when electronic warfare UAV load optimization and a variety of typical scenarios route planning recommendations.

With the development of aviation industry around the world, unmanned aerial vehicles (UAVs) are broadly applied in modern warfare and civilian fields, which induces a big boom in the number of aircrafts, thus posing a great threat for flight security with the increasing aircraft density in the sky in the near future. Although current researches and methods can deal with the UAV safety problems well at some extent, the effectiveness of them is greatly reduced when the UAV is in a highly antagonism and harsh environment. In this work, the autonomous safety control method for the UAV system is researched and explored from the need of improving the operational security of UAV system. UAV safety control method based on cognition-guidance refers to that the UAV can rely on perceptual and cognitive computing functions of the UAV system during the flight guidance to perceive and identify the forthcoming or ongoing security risks and to generate a correct evasion strategy. Cognition-guidance is a specific application of the cognitive computing. The safety connotation of the UAV and the research status of safety control are firstly analyzed, then, the relationship between the guidance and human cognition is discussed, and the UAV safety control method based on cognition-guidance is proposed. Besides, the UAV anti-collision control strategy is designed, and taking the two UAVs anti-collision control mission as an example, the simulation research for the UAV safety control method based on cognition-guidance is studied. And the simulation results have proved that the method is effective.

In the first place, this paper has introduced the types of domestic and international main current combat intelligent unmanned aerial vehicle. Secondly, the main system composition and function of unmanned aerial vehicle is presented. Additionally, the key technologies of unmanned aerial vehicle are discussed respectively, including autonomous control technology, path planning technology, navigation and positioning technology. Moreover, the related algorithms of unmanned aerial vehicle are summarized, including the PID control algorithm and fuzzy controller for flight control, the Kalman filter algorithm based on the analysis of gyroscope, accelerometer and magnetometer for improving the accuracy of attitude information, group intelligent optimization algorithm, the bionic intelligent optimization algorithm and the Neural network intelligent algorithm based on machine learning for unmanned aerial vehicle route planning,. Finally, the research areas are proposed to address development tendency and challenges of combat intelligent unmanned aerial vehicle.

Unmanned aerial vehicle (UAV) today play an increasing role in many public missions such as border surveillance, wildlife surveys, military training, weather monitoring, local law enforcement and so on. Due to its advantages such as low life-cycle costs and fewer restrictions on use, UAV has become important military weapon, then Counter-Unmanned Aerial Vehicle (CUAV) technology also become a research focus. Small drones that are not easily detectable have become an emerging threat. Low-cost and effective counter-UAV methods are explored for many countries. The CUAV technologies include traditional air defense weapon system, electronic warfare, UAV, laser weapons and so on. With the development of high-power lasers and continuous advances in beam control technology, CUAV laser weapons will become more practical.

Under the requirements of flexible, stealthy and miniaturize, the energy system has become the most critical factor that constraint the performance of smart unmanned equipment. Due to its advantages of convenient movement and strong environmental adaptability, the sphere platform shows broad prospects in military and civilian application. This paper designed an energy self-sufficient sphere platform equipped with intelligence reconnaissance device. The energy module of the device consists of an inscribed polyhedron solar cell array. The high power density energy storage component inside the sphere realizes energy storage and power supply for the signal transmission module. The sphere is equipped with an autonomous mobile unit and a detection unit, activated by outside trigger signal, and can be used in close reconnaissance, short-range detection and strategic destruction. The output power of the prototype machine produced by 3D printing is close to the theoretical result, and it maintained in high value during device rotation, laying the technical foundation for the practical application of this spherical platform.

Based on the advanced UAS GCS at home and abroad, the project studies the modular analysis of its functions and composition, demonstrates and studies the key technologies and realization methods of the large GCS simulation system. The UAS simulation training and teaching system has established a realistic command and control simulation training environment. It has mission planning, real-time visual display and teaching management functions, and can complete various training and operation procedures for UAS. It can greatly improve the training efficiency of the UAS operator and shorten the training time of the personnel.

The simulation system for Multi-UAV cooperative target tracking is an important supporting tool, needed by theoretical verification, algorithm experiment and 3D real-time visual simulation. Firstly, based on the HLA standard, the 3D simulation system is designed through functional orientation, structure design and operation process management. Then on the basis of flight dynamics of UAV, Aerosim Toolbox is used for simulation system. Meanwhile a method of realtime and seamless rendering of infinite terrain is implemented by block duplication, and the management of the detail scene is realized with BSP method. Thus the key technology of the simulation system is built. If users understand the relevant interface protocol, they can design their own algorithms and do experiments on the simulation platform. Finally, the platform is used to simulate the formation control algorithm and Multi-UAV target tracking flight control algorithm. The simulation results show that the system has strong generality and fidelity.