NCD ( Nonlinear Complex Diffusion) can adjust the diffusion intensity according to different structural features and effectively suppress the random noise in seismic data． However，the morphology of desert seismic data is complex，so that the NCD method is unable to effectively estimate the direction of the reflection event，leading to the attenuation of the reflection event． Aiming at this problem，an ASOCD ( Adaptive Structure-Oriented Complex Diffusion) algorithm is proposed based on the structural features of the reflection event．The oriented complex diffusion is performed along the direction of refection event to suppress the random noise and enhance the effective signal． And the proposed algorithm adaptively adjusts the diffusion threshold according
to the regional characteristics of seismic data，thus better recovers the seismic signal using a small diffusion threshold for the signal region． The results of the synthetic and real desert seismic data verify the effectiveness of the proposed method，and the proposed method can increase the SNＲ ( Signal-To-Noise Ｒatio) of filtered data by 6 dB more than the NCD method．

As a new modulation method，high-dimensional modulation solves the contradiction between spectral efficiency and power efficiency． However， the high complexity of lattice coding and decoding result in transmission delay，which seriously affects the application of high-dimensional modulation technology in highspeed and high-capacity optical communication systems． To solve this problem， a low complexity high dimensional lattice coding and decoding method based on tree structure is proposed． The tree structure is constructed according to the lattice distribution law，and the lattice coding and decoding is finished by the tree
structure index algorithm． This method can effectively realize high-dimensional lattice coding and decoding．When the SNＲ is 16 dB，the bit error rate can reach 3． 1 × 10 － 4，and the complexity of lattice coding and decoding with points of M can be reduced from O( M) to O( log M) ，and is universal．

To further strengthen the robust control of Lurie system，based on the early theoretical study，a controller design theory for Lurie systems with time-varying delay uncertainty is discussed with the application of the new cross-term definition method． By constructing a new Lyapunov functional and following the characteristics of Lurie system，the new feedback controllers are designed for singularly perturbed Lurie systems with uncertain time-varying delay in the sector region ［V1，V2］，and the state feedback control laws of time-delay dependent and time-delay independent are deduced． Finally，the feasibility and effectiveness are illustrated by the correspanding numerical example．

The jacket platform is frequently used platforms in the development of marine oil and gas，but its marine environment makes the jacket often subjected to fluid erosion and seawater corrosion，so it is necessary to check and maintain regularly． In order to shorten the walking distance of ＲOV( Ｒemotely Operated Vehicle) ，reduce the cost of detection and maintenance and improve the intelligence level of path planning，a jacket platform in Huizhou Oilfield was selected as the research object． In view of the two detection conditions of the single side and the whole detection of the jacket by ＲOV，the path planning is carried out in combination with the theory of graph theory and combinatorial optimization． The full traverse route of the jacket is obtained through simulation analysis，path planning method and simulation results have great theoretical value and practical
significance．

In gas pipeline leak detection，small leaks are most difficult to detect． And the main difficulties are noise reduction and feature extraction． VMD ( Variational Mode Decomposition ) is an adaptive signal decomposition algorithm with good anti-noise performance and the ability to extract the center frequency of the signal． A VMD improved algorithm is used to detect small leaks． The error energy algorithm is used to screen the IMF of the signal after the variational mode decomposition to obtain the final reconstructed signal． In order to extract the leakage components in the acoustic signal of gas pipeline，a feature extraction method combining VMD and error energy algorithm is proposed． Firstly use this method to decompose the original signal into a series of bandwidth-limited natural modes; then use the error energy algorithm to select the useful mode; finally，use the selected mode to perform signal reconstruction． The experimental results show that the method can effectively filter out the noise in the small leakage signal and effectively reconstruct the original signal．

To solve the problem of real-time monitoring and harmonic analysis of current signals in circuits，a high precision mutual inductance current signal detection device based on FFT ( Fast Fourier Transform) is designed by using sensor technology． The device achieves the purpose of non-contact measurement of loop current signal． It consists of hardware and software． The hardware circuit includes mutual inductance current sensor and detection circuit． The STM32 is used as the main controller performing，Fourier transform to get the frequency and amplitude information of the signal． This microprocessor-based measuring device realizes measurement automation and function diversification，and has certain application value in the industrial field．After detection，the frequency resolution and the amplitude resolution achieve 0． 5 mA，the frequency error of the
current signal is less than 2%，and the amplitude error is less than 5%．

The properties of gas sensors prepared in the experimental course do not satisfy the need for practical application due to the dated content and method of experiment used． To solve that problem and meet the need that the contents of experiments should relate to scientific research，the experiment of preparing and gas sensing properties of semiconductor gas sensing material are studied to enhance the properties of gas sensors． Ternary metal oxide Zn2SnO4 was prepared by solvothermal method and the sensing properties were measured systemically． This experiment content was applied as an open innovative experiment for the course of electronic functional material experiment． The results show that the prepared Zn2SnO4 based sensors exhibited excellent response to formaldehyde． Through the practical process，the students mastered a new method of preparing
nanomaterials and learned the characterization methods． The application result demonstrates that this experiment can help developing scientific research quality and innovation ability．

To solve the problem of semiconductor doping concentrations，it is necessary to test the semiconductor in a simple and easy way． Four-probe method is a common method of measuring resistivity in the field of microelectronics technology． Combining the construction of the semiconductor device physics and experiment course，the experiment of testing the doping concentration of semiconductor by four-probe method was developed． Through the establishment of two theoretical models of semi-infinite sample model and infinite thin-layer sample model，the resistivity test methods for semiconductor materials with different thicknesses are studied，and the principle is discussed． In order to solve the problem that the commercial equipment is expensive and can not meet the needs of experimental，the author proposed to build a test system，using a simple manual probe station of tungsten alloy，and designed a four-probe test device according to the experimental needs． The practical application shows that the establishment of the four-probe test system meets the teaching requirements of semiconductor physics experiments．

In order to help learners master their individual characteristics and learning information defined by adaptive learning system，an open learner moder，called as MindOLM，has been designed and developed．MindOLM adopts mind map to release curriculum knowledge states in learner model to the learner，helping the learner to understand and view his /her learning progress，and the relationships among chapter，section and knowledge point，and inducing his /her meta-cognitive learning experience． In the empirical research，evaluation data was collected through experimental research method，questionnaire survey and face-to-face interviews． The results show that: 1) MindOLM can induces meta-cognitive learning experience on learning planning and self
reflection; 2) MindOLM can presents their learning progress and knowledge level; 3) learners have shown strong acceptance to usability，usefulness，and satisfaction with MindOLM．

Seismic signal is a kind of non-stationary signal，which is composed of seismic subwaves with different amplitude and frequency at different propagation times． According to the time-frequency characteristics of seismic signals，a denoising method based on the relevant energy entropy threshold of variational modal decomposition is proposed． The variational modal decomposition algorithm is used to decompose seismic signals into modal components from high frequency to low frequency and a certain bandwidth． The normalized correlation coefficient between the modal components and the seismic signal is calculated． And the effective information and noise in the modal components are located． Each modal component is divided into several sub-ranges，and the noise energy entropy of each sub-interval is calculated separately． The modal component coefficient of the maximum range of energy entropy is selected as the noise variance of the component to obtain the threshold of the component． The synthetic seismic model and the actual seismic signal are de-noised and compared with the variational modal decomposition and de-noising method that directly removes high-frequency components． The results show that this method can effectively extract the active components of seismic signals and improve the signal-to-noise ratio．

In order to solve the problems of fragmentation of knowledge and lack of semantic relationship in digital network courses，to deepen the control unit of knowledge to the level of knowledge element，the finegrained curriculum knowledge element organization model is constructed in combination with CELTS standard．And then use the theory of fine processing to arrange and design the teaching content，and define the logical relationship among knowledge element，knowledge chain and knowledge unit． The knowledge graph of C course is constructed to visualize the granularity of knowledge elements and the relationship between knowledge elements through Neo4j． Practice shows that the combination of knowledge elements and knowledge graph can effectively represent the logical relationship between knowledge and improve the correlation between knowledge．

In order to solve the problem that the traditional access control system has low security and poor user experience and can not meet the social demand for smart home，an artificial intelligence access control system based on remote terminal controllable is developed． The system is based on the Intel UP2 embedded platform．The Ubuntu operating system is the software foundation． It integrates technologies such as face recognition，fingerprint recognition，voice recognition，and ＲF card recognition． Write a graphical interface in PyQt language and builds a cloud server to implement remote terminals control． The system integrates a variety of identification and unlocking methods，which can realize the intelligent and convenient unlocking，and better adapt to different life situations．

In accordance with the requirements of the analog circuit experimental teaching system，in order to help students to further master the analog circuit signal generation and processing method，by abstracting specific circuit elements into mathematical models，an analog circuit signal generation and processing teaching system is developed based on LabVIEW software． According to the content of the teaching experiment，three experimental modules are designed． Each module contains different experimental items，covering the basic part and actual application of analog circuit signal generation and processing． The design principle of the experimental module is given，and the experimental principle and content of the different modules are discussed． The specific experimental case module is taken as an example to explain the design and use of the lock-in amplifier． This system integrates the basic knowledge of signal generation and processing of analog circuit and expands the application of lock-in amplifier，instrumental amplifier and linear transformation． The practical application shows that the experimental system based on LabVIEW meets the experimental teaching requirements of the course and has achieved good teaching results．

To evaluate and select a signal remote Ｒemo ( LEMO ) connector for ITEＲ ( International Theramonuclear Experimental Ｒeactor) soft X-ray diagnosis conveniently and quickly，based on the threedimensional electromagnetic simulation software CST ( Computer Simulation Technology) ，a three-dimensional model of a connector ( socket ECG． 3B． 332． CLV and plug FGG． 3B． 332． CLAD12) is established． The effects of transmission delay，scatter parameter，pin core corner value and crosstalk on transmission signal integrity are simulated and analyzed，and compared with the test results of the crosstalk of the transmission signal of the connectors by spectrum analyzer and other devices． The simulation results show that the insertion loss is not less than － 0． 5 dB． The insertion loss and return loss corresponding to the corner value of the pin core have exponential relationship with the frequency，and the crosstalk amplitude has quadratic function relationship with the frequency in the working frequency band of the Ｒemo connector． The simulation results are in good
agreement with the crosstalk test results． The simulation method and research findings can be used to test and evaluate the reliability of other diagnostic signal connectors in the subsequent plasma diagnostic environment．

In order to solve the problem of insufficient practical operation of students in the traditional teaching process of asphalt mixture rutting experiment，the asphalt mixture rutting virtual simulation system is developed．Based on Unity 3D virtual reality platform，and virtual reality technology is introduced． Virtual simulation and interaction design are carried out according to real experimental environment and process in this system，which has four modules: experiment guidance，equipment recognition，virtual simulation and experiment report． The operation mode includes automatic demonstration and human-computer interaction for students to choose freely，which realizes 3D virtual simulation experience learning of asphalt mixture rutting test，and its running effect is good． The system solves the difficulties in undergraduate teaching of asphalt mixture rutting experiment，which is helpful to improve the efficiency of experimental teaching projects，and is of great significance to cultivate students' practical ability and operational ability．

Traditional intrusion monitoring system can not keep up with the speed of malicious code upgrade，monitoring effect is poor． Therefore，an automatic security monitoring system for malicious code intrusion in local area network is designed． In the hardware part of the system，the malicious code packets with intrusion and attack behaviors are stored and matched by the database module，and the restored information is stored． The features of the packets with intrusion are extracted by the intrusion monitoring module，and the intrusion behaviors are judged． Through the database restore module to the network transmission packet interception，and restore the data processing and storage; Through the log audit module，the information of intrusion into the system is transmitted to the database． On this basis，a clustering algorithm is used to effectively monitor the malicious code invading the system． The experimental results show that the monitoring system has the advantages of high coverage monitoring rate，less monitoring time and low error rate．

To assist doctors in detecting large numbers of pulmonary nodules in lung CT ( Computed Tomography) images，lung nodules were segmented using the deep U-Net network． As U-Net network comes from FCN ( Full Convolution Network) ，it is characterized by end-to-end and pixel to pixel． The network classifies each pixel and can combine the abstract feature information of deep features with the position information of the concrete feature of shallow features to obtain the segmentation mask of CT image． Experimental results show that this method is effective in detecting lung nodules．

In order to make undergraduates to master the remote sensing image processing technology，and combination with college student innovation training program in Jilin University，a land classification project of campus based on high-resolution satellite remote sensing image is designed． In this paper，GF-2 satellite image data of the central campus of Jilin University on September 20，2015 is used as experimental data． Spectral index and SVM ( Support Vector Machines) are used to get the buildings，vegetation，water，bare land and playground of the campus based on high resolution remote sensing image． In addition，the classification result is analyzed statistically to get campus vegetation coverage，campus building area，and land area to be developed． What’s more，the statistical results are compared with ground truth data to improve the classification accuracy． The experimental results show that the experiment can effectively help students to master knowledge of remote sensing，and get the expected teaching result．