Aiming to develop terahertz multi-band high sensitivity receiver, a broadband 0.67 THz sub-harmonic mixer is developed based on the nonlinearity of terahertz GaAs plannar Schottky diode and quartz film thinning process. The influence of the performance parameters of Schottky diode on the performance of terahertz mixer is analyzed, which provides a theoretical guide to diode selection and mixer evaluation. In the design of 0.67 THz sub-harmonic mixer, the synthetic design method is adopted by using the electrical and electromagnetic simulation software, and the optimization is made on the coupling between the discontinuous microstrip and waveguide in order to minimize the conversion loss. In the design of 0.67 THz sub-harmonic mixer, the best simulated conversion loss is less than 7 dB over 0.62-0.72 THz frequency range, the LO drive power is less than 4 mW, and the isolations are above -30 dB between LO port and IF port, RF port and IF port.

The low Terahertz(THz) radiation power has become the main bottleneck restricting the development of Terahertz science and technology. In order to increase the power of THz source, a novel method of detached cavity is proposed to multiply frequency based on double drift region avalanche multiplier diodes, and then the power combiner of waveguide is designed and realized. The phase of the intermediate frequency of the excitation source is adjusted on the input of the frequency multiplier. And then, the phase of the frequency multiplier output radio source can be controlled precisely. Sequentially, the independent signals of THz which are in-phase can be added together. The efficiency of the power combiner is enhanced avoiding using the phase shifter of THz frequency band. The complexity of the processing is reduced and it is easy to be realized. The simulation results and the measured data show that, the efficiency of the power combining working at THz band is 79%-89% approximately.

In order to meet the requirements of the terahertz wireless communication system for large-capacity baseband signal processing algorithms, basing on the traditional filter parallel implementation algorithm derived directly from polynomial decomposition, a parallel implementation of a fast Finite Impulse Response(FIR) filter with less complexity is derived by matrix variation. On this basis, the 2 parallel, 4 parallel and 8 parallel conversion formulas and implementation architecture are given by the tensor product representation. The general implementation formula of the parallel fast FIR filter is derived and the complexity difference before and after optimization is compared. Finally, the derivation formula and specific implementation architecture of the 64 parallel fast FIR filter are given, as well as the hardware complexity comparison before and after optimization.

：The effective second-order nonlinearity and terahertz emission properties of DAST crystal are mainly studied. Based on the metastable region of DAST-methanol solution, the growth experiment of DAST is carried out by solution cooling methods. It is found that the higher the cooling rate, the faster the crystal growth rate, but the polycrystalline is easy to occur. In the later stage of crystal growth, the lower cooling rate is beneficial to the increase of crystal thickness. After grinding, the surface roughness of the polished crystal could meet the optical test level(micron level) requirement. The average effective second order non-linear coefficient of DAST chip is 16.58 pm/V. Terahertz emission in the frequency range of 0.84-10 THz is realized, and the maximum emission intensity is at 2.72 THz.

At present, Time Division Multiplexing-Multiple Input Multiple Output(TDM-MIMO) technology is mainly adopted to increase the number of virtual antennas to improve the angular resolution in automotive radar. However, when there exists relative motion between the target and the radar, the amount of phase change caused by the Doppler frequency of the moving target during the switching time of different transmitting antennas is coupled to each receiving antenna, resulting in a defocusing effect of the spectrum. This phase change would have effect on the angle estimate of the target. A TDM-MIMO signal model is deduced in this paper. By analyzing the cause of the phase error of the moving target, a phase compensation method is proposed. It is unnecessary for this method to estimate the target speed, and no additional hardware overhead is needed. The simulation experiment and the measured data show that the method can realize normal angle measurement on the moving target with low time complexity.

The 3D massive Multiple Input Multiple Output(MIMO) under the Rician channel model is studied, and the easily-implemented transmission scheme with high-spectral efficiency is proposed. The proposed scheme only utilizes the Line-Of-Sight(LOS) channel, as it is easier to be obtained than the scatter channel. Firstly, the user scheduling algorithm is utilized to solve the problem of too much interference between adjacent users, so as to improve the sum rate of the system, and then the precoding method of domain selection is utilized to eliminate the interference between different users and maximize the Signal-to-Noise Ratio(SNR), so as to further improve the system sum rate. The simulation results validate that the proposed scheme has higher spectral efficiency than the previous schemes.

The outage performance of cooperative Non-Orthogonal Multiple Access(NOMA) network adapting Amplify and Forward(AF) is investigated. In the multi-user multi-relay cooperative NOMA network, in order to guarantee the Quality of Service(QoS) requirement of the signal with low latency in the mixed signal, a power allocation scheme is proposed, and then the detection constraint conditions are obtained. Furthermore, for the given relay, the set is built according to the detection constraint conditions, and the user relay pair which can maximize the rate of high QoS signal is selected as the best “user-relay” to transmit the message. The expression equation of the system’s outage probability for the proposed user-relay selective scheme is derived and the asymptotic expression is solved. Simulation results prove the correctness of the derived results, and show that the system’s outage probability decreases with the increasing of the number of the relays. Especially when the Signal-to-noise ratio tends to be infinite, the system’s diversity gain is depended on the number of the relay. In addition, the outage performance of the proposed scheme is better than that in other existing literatures.

The communication countermeasure simulation based on Multi-Agent is studied. According to the battlefield environment, actual needs and the synergy between Multi-Agent, the attributes, parameters and behaviors of Multi-Agents are analyzed and defined. According to the design rules of Observe-Orient-Decide-Act(OODA), a four-Quadrant Attack(QA) strategy for maximizing the individual value of the agent is proposed to achieve the autonomous attack of agents. The simulation is implemented by Repast Simphony modeling software. The effect is comprehensively observed through 2D model simulation and the effectiveness of the algorithm is backtracked. The proportion that can maximize the effectiveness of Red Attacker Agent(RAA) attacks can be tested according to the number of Blue Communication Agent(BCA) attacks. Practical application testing shows that this model has high attack rate, high realization value, which provides a new idea for the simulation of communication antagonism.

Melbourne-Wbbena(MW) combination method cannot detect the same cycle slip in the same epoch of L1 and L2 carriers, and the cycle slip separation cannot be realized after the sounding of the cycle. An improved method of cycle slip detection based on jointing Doppler and MW combination is proposed. The Doppler integral is applied to detect and repair the cycle slip of single frequency point. Combining the traditional MW method and Doppler integral method, an improved model is established, and the cycle slips generated by L1 and L2 carrier phases are calculated and separated. Verification and analysis are carried out by the measured data, and the detection results of the improved methods under three different conditions are compared. Experiments show that the new algorithm can not only detect and separate the cycle slip combinations in L1 and L2 carrier phases accurately, but also accurately detect the location and the values of cycle slips; and it can detect and repair the cycle slips for more than 1 cycle, which effectively improves the accuracy and reliability of the MW combination algorithm.

Frequency Diverse Array Multiple-Input and Multiple-Output(FDA-MIMO) has drawn a remarkable amount of attention due to its particular range-dependent beampattern. Firstly, the echo model of the FDA-MIMO radar is established; then the conventional Eigensapce-Based(ESB) beamformer is extended to this model. Finally, an improved ESB beamforming algorithm is proposed to enhance the performance in low Signal Noise Ratio(SNR). Simulation results show that the range-dependent interference from main lobe can be suppressed and the effectiveness of the proposed algorithm is verified.

Typhoon is a disastrous weather system which has a great change of path and intensity with ephemeral and strong collapsing force. Satellite remote sensing, especially satellite microwave imaging sensors, not only have the advantages of large scaled, dynamic, synchronous, fast observation etc., but also can work 24 hours all-weather, which can be used for monitoring and early warning of typhoon path and thus is considered as an efficient tool to analyze typhoon. Based on the data of three independent sensors including microwave satellite Advanced Microwave Scanning Radiometer(AMSR), TOPEX&Jason-1 altimeter and WindSat radiometer, the influence of typhoon on ocean is analyzed. The satellite measurement of ocean parameters relevant to typhoon is described, and finally the comparison of wind field ocean remote sensing results with currently used radiometers and scatterometers is shown. The conclusions are given as following: compared with infra radiometer, microwave radiometers have an observable advantage; and full polarization radiometers have the potential of measuring ocean wind field in bad weather conditions.

Aiming at the problem of node location in Wireless Sensor Network(WSNs), Unmanned Aerial Vehicle-assisted localization algorithm based on Feedforward Neural Network(UAV-NN) is proposed. The localization is performed by using mobile Unmanned Aerial Vehicles(UAVs) as the anchor nodes to send the beacon signals every period of time, thus every unknown node can estimate its current position based on the Received Signal Strength Indicator(RSSI) values of the received beacon signals by training the Single hidden-Layer Feedforward Neural Network(SLFN) using Extreme Learning Machine(ELM) technique. The proposed method requires fewer anchor nodes and no ground anchor node compared to traditional RSSI based localization technique to yield better accuracy. Simulation results show that this technique is capable of performing real-time localization for unknown nodes with less localization error by using ELM compared to other traditional machine learning algorithms.

Data transmission is the basis of Vehicular Ad-Hoc Networks(VANETs) for traffic safety. The movement of vehicles, the attenuation of signal transmission and the interference between each other have important effects on the reliability of the link. The Dynamic Transmission range-based Multi-hop Stable(DTMS) routing is proposed, which takes into account the reduced transmission distance due to attenuation when selecting the next-hop forwarding node. Firstly, the dynamic transmission distance of the vehicle is estimated, and then the connection time and distance rate of the link are estimated based on the dynamic transmission distance. The weight of the neighbor node is calculated. A timer is set based on the weight information of the node to generate the next-hop forwarding node. Simulation results show that the proposed DTMS routing can effectively improve the success rate of packet transmission.

In order to solve the problem of node coverage suppression caused by blind area coverage and frequent data retransmissions in the pre-coverage process of wireless sensor networks, a coverage algorithm for wireless sensor networks based on reserve satisfaction model is proposed. Firstly, a new wireless sensor network node coverage model is designed, and the coverage index, coverage intensity, coverage balance evaluation coefficient and other evaluation dimensions are constructed to quickly evaluate the coverage quality of nodes. Secondly, the coverage equilibrium evaluation coefficient is calculated, and the backup model is adopted to optimize coverage quality and determine the backup working nodes with superior coverage performance. Then, based on the principle of coverage similarity, the coverage performance of working nodes is evaluated, and the first coverage evaluation method is designed. The statistical mean of coverage index is evaluated according to the moving path of the nodes. According to the sequence of the target nodes entering the coverage area, the working nodes with the best performance are activated. The simulation results show that the proposed scheme has higher network coverage, shorter coverage start-up time and less work compared with the current constant node coverage schemes of wireless sensor networks.

The spaceborne low Passive Intermodulation(PIM) feed is designed for some type of high-power feed product. The simulation analysis method and the test results are given in detail. The proposed design utilizes choke flange with low PIM to depress the conduction current with third-order intermodulation frequency on the joint face in order to reduce the product of PIM, and the PIM value can be decreased by more than 40 dB. To decrease PIM for other non-third-order frequencies, the pressure of flange can be enhanced. The performance is compared with those of similar foreign products, which fully shows the validity of the design and its technical level.

Microwave and optical measurements are the best diagnostic methods for electron-beam-generated air plasma. According to the microwave characteristic of plasma, the attenuation in energy and phase shift can be adopted to diagnose electron density. The experimental results of microwave and optical measurements are coincident. From the experiment results, humidity effect on air plasma is analyzed, and it is found that the electron density of humid air plasma is obviously reduced compared to that of the dry air plasma under the same experiment conditions. The numerical calculations using the chemical processes model are completed for air plasma, and the numerical result is in a good agreement with the experimental result.

The electromagnetic security of datalink becomes an outstanding issue for restricting the development of an Unmanned Aerial Vehicle(UAV). To improve the adaptability of the UAV’s datalink in the electromagnetic environment, a test method of electromagnetic radiation effects on the dynamic datalink is designed due to variable electromagnetic susceptibility depending on the flight state. Then the test of continuous wave radiation effects on the datalink is conducted. The results show that the electromagnetic susceptibility depends on the operation signal power. The electric field thresholds of the in-band interference is lower than 1 V/m under normal operations for UAV, but the corresponding Automatic Gain Control(AGC) voltage is the largest, reaching 190 V. Although the anti-interference capability of the datalink is weak, but its error tolerance performance is good. The increasing trend of the number of bit errors with the increase of Electromagnetic Interference(EMI) signal power is obvious. When the amount of bit errors approaches the limited maximum, the uplink can still work properly. Besides, the frequency offset EMI may suppress the BER increasing process at the critical lost-link state resulting in the disruption of the datalink. The corresponding susceptibility threshold becomes larger as the frequency offset increases, but the corresponding AGC voltage is relatively reduced.

As a new type of clean energy, biodiesel has been paid more and more attention. According to the existing methods of preparing biodiesel by chemical method, there are many disadvantages such as long reaction time, high energy consumption and serious environmental pollution. A new type of continuous flow microwave reactor is designed for the preparation of biodiesel. The electromagnetic field, chemical reaction kinetics, laminar flow and heat and mass transfer are coupled with multiple physical fields to simulate the reaction process through Coｍsol Multiphysics. By analyzing the effect of reactant flow rate on temperature, and by analyzing the effect of coil turns, coil spiral size and cavity size on the energy utilization ratio, the overall reactor structure size is optimized to improve the energy utilization ratio and increase the handling capacity of the reaction materials.S参INGH S P,SINGH D. Biodiesel production through the use of different sources and characterization of oils and their esters as the substitute of diesel:a review[J]. Renewable and Sustainable Energy Reviews, 2010,14(1):200-216.魏明红,赵华.生物柴油制备方法及应用现状[J].当代化工, 2006,35(4):246-249. (WEI Minghong,ZHAO Hua. Preparation method and application status of biodiesel[J]. Contemporary Chemical Industry, 2006,35(4):246-249.)袁红,刘慧昌.微波强化固体酸催化酯化反应[J].中国油脂, 2013,38(9):60-63. (YUAN Hong,LIU Huichang. Solid acid catalyzed esterification strengthened by microwave[J]. China Oils and Fats, 2013,38(9):60-63.)王如意,徐会佳,茹世杰,等.磷酸银光催化降解甲磺酸吉米沙星的研究[J]. 太原科技大学学报, 2017,38(3):243-248. (WANG Ruyi,XU Huijia,RU Shijie,et al. Simulation of constant velocity flow of pure material dendritic growth using phase field method[J]. Journal of Taiyuan University of Science and Technology, 2017,38(3):243-248.)张增强,孙楠,高锦明,等.微波辅助棕榈油制备生物柴油的研究[J].中国油脂, 2008,33(3):53-55. (ZHANG Zengqiang,SUN Nan,GAO Jinming,et al. Preparation of biodiesel from palm oil assisted by microwave[J]. China Oils and Fats, 2008,33(3):53-55.)韩毅,邓宇,郝敬梅,等.NaOH催化微波法制备生物柴油的工艺研究[J]. 精细石油化工进展, 2008,9(1):38-41. (HAN Yi,DENG Yu,HAO Jingmei,et al. Research on preparation of biodiesel by NaOH base catalytic under microwave irradiation[J]. Advances in Fine Petrochemicals, 2008,9(1):38-41.)

Compact Antenna Test Range(CATR) is an effective alternative method to measure the antenna pattern, however, the existence of the stray signals would significantly degrade the measurement accuracy. The paper proposes an effective method which can effectively eliminate the influence of the stray signals. The proposed method measures the Antenna Under Test(AUT) repeatedly while moving the AUT along any straight line with equidistance. There are different phase factors for each stray signal at different displacements. A Vandermonde matrix can be constructed by using these phase factors. All responses, including the desired response, can be calculated by inverting the Vandermonde matrix. The deviations between the simulations and test data are less than 0.5 dB, which validates the effectiveness of the proposed method.

In order to improve the resolution of low-resolution infrared images, this paper proposes an infrared image super-resolution algorithm based on random forest. Firstly, two random forests models are trained independently. The first model is trained by using infrared images, while the second is trained by using registered multi-sensor images. Then, an adaptive extraction algorithm is utilized to extract the edges of infrared images and registered visible light images. The correlation coefficient between the low-resolution patch of infrared image and the high-resolution patch of visible light image is calculated. According to the correlation coefficient, an appropriate model can be selected. Finally, the selected model is utilized to reconstruct the high-resolution infrared patch. All these patches are integrated into a high-resolution infrared image. The experimental results show that the proposed method can obtain better performance compared with the super-resolution random forest algorithm. The Peak Signal to Noise Ratio(PSNR) of testing images is increased by 0.09 dB on average. The reconstructed images, with better visual effect, are closer to original images.

Current image matching algorithms mainly use the distance information between pixels to achieve feature matching, ignoring the variance information between images, resulting in more false matching in the matching results. An image matching method is proposed based on Laplacian feature constrained coupling variance measure. Firstly, Harris operator is introduced to extract image features roughly. On the basis of rough extraction, Laplacian feature of pixels is utilized to optimize the extracted image features in order to obtain more accurate image features. Then, the gradient feature of the image is employed to calculate the direction information of the image. Based on the gradient feature, the neighborhood of the feature points is established, and the Haar wavelet value in the neighborhood is solved to obtain the feature vector. Finally, the regional variance model is adopted to measure the variance information of the image, and it is introduced into the process of image feature matching. The variance information is added on the basis of Euclidean distance measurement of feature points to achieve image feature matching more accurately. Random Sample Consensus(RANSAC) method is adopted to purify the results of feature matching, eliminate mismatching and complete image matching. The experimental results show that compared with the existing matching algorithms, the proposed algorithm has better matching performance and higher accuracy, with accuracy above 90%.

In order to solve the problems like false detection and missing detection in the recognition results due to ignoring the relationship between different color components when using the image forgery detection algorithm to locate tampered content, the image forgery detection algorithm based on multivariate exponential moments and Euclidean locally sensitive Hash is proposed. The Gaussian low-pass filtering is introduced to eliminate the noise in suspicious images. Subsequently, the filtered image is segmented into a series of overlapping circular sub-blocks for improving its robustness to rotation and other content operations. Based on the Quaternion Exponential Moment(QEM), each circular sub-block is computed to extract the corresponding robust features and combine them into feature vectors. The corresponding Hash sequence of each sub-block is generated by the Euclidean locally sensitive hashing mechanism. The spatial distance between any two adjacent Hash elements is computed, and the matching of all sub-blocks is finished by comparing with the preset threshold. Finally, by means of the consistency method of random samples, the false matching is eliminated, and the tampered content is located by morphological processing. The experimental results show that compared with the existing forgery detection methods, the proposed algorithm has higher accuracy of forgery detection under various geometric modifications.

Based on the nondestructive identification requirements of rice varieties, the spectral image features of three rice samples are analyzed by hyperspectral technique, and the detection, classification and identification of three kinds of rice using Liquid Crystal Tunable Filter(LCTF) spectral camera are realized. The VIS/NIR(Visible/Near-Infrared) spectral images of rice samples are collected by hyperspectral camera, and the hyperspectral image is processed and analyzed by Matlab and ENVI software. The relative reflectance curves of each sample are obtained. By using image threshold segmentation technology, the spectral images of each band are obtained. Combining the images and data, the spectral differences of different varieties of rice are analyzed. It is found that the rice had a distinct characteristic peak in the 480-550 nm band. The spectral differences between different varieties are obvious, and the ratios of the brightness of the binary images for different varieties of rice are different as well. The results show that the relative reflectivity and binary image of spectral images have good prospects in the application of rapid classification and identification of rice varieties.

As a noninvasive brain imaging method, structural Magnetic Resonance Imaging(sMRI) plays an important role in understanding the pathology of Alzheimer’s Disease(AD). Convergence evidence suggests that imaging features based on sMRI performs well in classifying AD patients from Normal Controls(NC). In this paper, a simple feature extraction method is proposed based on the Human Brainnetome Atlas with 227 AD patients and 226 NC from Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Firstly, mean gray matter density of each brain region in the brainnetome atlas is obtained as feature. After that, an Support Vector Machine(SVM) model is introduced to classify AD from NC samples. The result shows that the mean accuracy is 85.2% with 10-fold cross validation. And post hoc analysis demonstrates that the mean gray matter density of several brain regions such as the hippocampus, amygdala and fusiform gyrus play important roles in classification and the atrophy of these regions has significant correlation with cognitive ability measured by the Mini-Mental State Examination(MMSE) scores in AD patients. In addition, the Least Absolute Shrinkage and Selection Operator(LASSO) is utilized to predict individual MMSE score and the predicted values show a very high positive correlation with true values (r0.65, p0.001). In conclusion, the present study demonstrates that the gray matter density of brain regions defined by the Brainnetome Atlas is meaningful for distinguishing AD patients from NC samples and could also be utilized to predict individual cognitive ability.

\ It is important to determine the genetic types of high risk of birth defects and to promote early heredopathia screening and birth guidance for first class prevention of birth defect like congenital hearing loss. This research applies the decision tree algorithm to analyze the clinical data of nearly 1 000 cases of GJB2 genetic mutation record by using data mining tools. The model of assistant screening for pathogenic genes of hearing loss birth defects is established. By studying the structure of the model tree and samples’ classification results, it is concluded that pure positive samples of hearing loss are obtained in five groups of branches in the model tree. Besides, the set of genetic loci states formed by each of these branches is consistent with the mutation state of pathogenic genes confirmed by clinical studies. The screening model established by the decision tree method can assist doctors to quickly screen out the types of pathogenic genes from the clinical big data.

In order to further analyze and predict the changing trends of the Fu Jian Ping Tan(FJPT) station's variable velocity field, and then grasp its motion variation characteristics, the stationary test is carried out based on the time series of the eastward speed and the northward speed of the FJPT station from 2010 to 2018. According to the test results, it is an unsteady random non-monotonic time series. Therefore, the gray system Verhulst model is employed to predict the eastward and northward annual average momentum of the next 10 years. The prediction results show that the speed of the FJPT station is gradually deflected from the west to the east and from the north to the south, and the station speed is deflected from the northwest to the southward. The south side of the southeast coastal block is in the southeast movement, and the FJPT station turns around the southeast coastal block in clockwise.

Joint control of speed and traffic timing based on Vehicular Ad Hoc Networks(VANETs) is one of the core technologies in intelligent transportation system. Therefore, VANETs-based Vehicle Speed Controlling Optimization(VSCO) scheme with traffic light is proposed. VSCO scheme optimizes the time of traffic lights(green, red and yellow) through VANETs, and plans the vehicles driving by the traffic lights to eliminate Yellow-Light Dilemma(YLD)-related traffic accidents occurring at intersections, reduce CO2 emissions and shorten the waiting time of vehicles. Simulation results illustrate that the proposed VSCO scheme can effectively decrease the waiting time and cut CO2 emissions, compared with conventional scheme referring to traditional traffic light using fixed signal control method.

In order to avoid the hotspot problem in a Wireless Sensor Network(WSN), a portable sink is utilized to gather information from the sensor nodes. A portable sink might be permitted to visit just a restricted number of sensor nodes, alluded as Rendezvous Points(RPs). The rest of the sensor nodes send their information to the closest RP so as to avoid the postponement due to visiting all the sensor nodes. Therefore, it is very important to determine an optimal set of RP in order to enhance the information gathering procedure. An Intelligent Water Droplet-based Sink Path Planning(IWD-SPP) algorithm is proposed. This algorithm aims at boosting up the lifetime of the network and minimizing the network energy consumption in forwarding the data packets. A productive mechanism has been devised based on the behavior of IWDs so as to find an ideal arrangement of RPs and the mobile sink's routing path. Simulation results show that the proposed IWD-SPP algorithm has better performance in energy consumption and network life.

Aiming for the load unbalance problem in Low Power and Lossy Networks(LLN) due to uneven node deployment, an Energy-Efficient Routing Protocol for LLN based on Load Balancing(EELB- RPL) is proposed. This protocol has made the following improvements：the optimal parent node is selected by comprehensive examination of the link quality, transmission delay, node residual energy, throughput and congestion detection factor. By adjusting the Trickle timer, the nodes adjust their suppression mechanism according to the network topology density, so as to avoid the suppression unfairness and load unbalance. The simulation results show that compared with the existing algorithms, the EELB-RPL can effectively achieve load balancing where the average packet delivery rate is increased by 14.6%, the throughput of sink node is increased by 28.5% and the network lifetime is increased by 8.96%.

Improved particle swarm algorithm is adopted in order to evaluate the flatness measurement error accurately. Firstly, mathematical model of flatness error evaluation is established. Secondly, particle swarm algorithm is improved, which is included inertia weight control based on membership function of double sigmoid type, Triangle function adjustment process, and selecting fitness function. Finally, the algorithm termination condition and flow are given. Experimental simulation results show that the convergence of improved particle swarm optimization algorithm is fast, the flatness error is 9.496 μm at an average of 30 experiments, which is smaller than that of other optimization; the standard deviation of the experiment is 0.048 2 μm, which is smaller than that of other algorithms as well, so that the evaluation precision is improved effectively.

The traditional Differential Optical Absorption Spectroscopy(DOAS) method which is accurate and fast, has simple detection principle for H2S gas concentration. However, big errors occur in environments with low concentrations and short optical paths. In this paper, inversion of low concentration H2S gas is executed by using genetic algorithm based on traditional differential absorption algorithm. But the genetic algorithm is prone to premature convergence and thus falls into local optimum. The genetic algorithm after disaster optimization is designed to invert the H2S gas concentration. The results show that the method has high measurement accuracy for low concentration H2S gas, and combines with traditional differential absorption spectroscopy to obtain a wider detection range.

The diffusion of proton in semiconductor electronic devices can cause performance degrading. Nevertheless, it is seldom known to date about the mechanism of this process in oxide. In this work, proton diffusion is studied in a-SiO2 using Ab Initio Molecular Dynamics(AIMD) from the atomic scale. The diffusivity, activation energy in a-SiO2 are calculated, and the diffusion paths are discovered. In a-SiO2, the proton is found to be diffusing via forming and dissociating bonds with oxygen atoms. Two diffusion means of proton in a-SiO2 are proposed, hopping diffusion and rotation diffusion. Meanwhile, the energy barriers for these two diffusion methods are calculated respectively.

The band gap width of ZnO/GaN heterojunction is large, which restricts absorption of visible light. In order to study the effects of Ag on the visible light absorption of ZnO/GaN heterojunction, GaN/ZnO heterojunction is constructed on the (1-100) nonpolar plane, and Ag is utilized to replace Zn and Ga atoms at different positions, respectively. Then, the influence of Ag doping on the stability, electronic structure, optical properties and band edge position of ZnO/GaN heterojunction is calculated by first principle. The results reveal that the system of Ag doped ZnO/GaN heterojunction is stable because of the negative formation energy. For Ag replaced Zn and Ga, the band gap width decreases from 2.93 eV to 2.7 eV and 2.3 eV, which results in red shift of absorption coefficient as well as photoconductivity, and is beneficial to the absorption of visible light. Finally, Ag doped ZnO/GaN heterojunction has excellent photocatalytic activity.

Lithium battery and its application have gradually become a research hotspot in recent years. In order to improve the estimation accuracy of Battery Management System(BMS) for battery State Of Charge(SOC) and State Of Health(SOH), an improved extended Kalman filter algorithm for online collaborative estimation of battery state of charge and health is proposed based on the establishment of the second-order Thevenin equivalent circuit model. The model parameters are obtained by the least squares method through a staged pulse discharge experiment. The performance of improved extended Kalman algorithm in SOC and SOH estimation accuracy, convergence of algorithm and complexity of algorithm under Dynamic Stress Test(DST) conditions are compared and analyzed by Matlab. Experiments show that the improved extended Kalman filter algorithm can accurately estimate SOC and SOH at each sampling point, with an error about 1%. Moreover, when the initial value is not accurate, the algorithm can rapidly converge to the true value with good robustness, and the accuracy of the algorithm is almost unaffected by the fine-tuning of model parameters.