In order to collect microseismic signals more accurately，ray tracing is used to simulate multi-source
microseismic signals． First，we used the two-point ray tracing method to adjust the initial value，and then the
optimal exit angle and corresponding ray trace are obtained，realizing full-wave ray tracing． Forward modeling of
microseismic signals based on ray tracing can accurately reflect the kinematic characteristics of stratigraphic
structures． It is a rapid and accurate forward modeling of microseismic signals method． By analyzing the regular
of seismic wave propagation，the forward model established can provide some help for signal analysis and
processing，inversion of microseismic data and source localization．

In order to solve the problem of passive synchronization control of fractional order Lorenz systems with
different dimensions，according to the stability theory and passive control theory of fractional-order system，a new
active controller and passive controller are designed to make the system realize the effect of synchronous tracking，
and to achieve passive synchronization of fractional-order systems with different dimensions under different initial
conditions． This method extends the passive synchronization control of integer order chaotic systems for the same
dimension and structure to the different dimension fractional order system with order less than 1． The numerical
simulation results verify the feasibility and effectiveness of the designed controller． The driving system can keep
the same trajectory with the state variables of the response system，and the control method is specific and fast．

In an island microgrid with multiple distributed generations operating in parallel，due to the difference
in impedance of each line，a reasonable control of reactive power can not be achieved using the droop control
strategy． This paper proposes a self-adjusting virtual impedance droop control strategy，which adjusts the virtual
impedance through reactive power and compensates the output voltage difference caused by the impedance
difference without detecting the line impedance parameters，so that the output reactive power of each inverter is
equally distributed or according to the capacity ratio． Using Matlab /Simulink，an isolated island microgrid
simulation model with two distributed generations and parallel operation was built． In two cases，it is verified that
the improved droop control strategy can achieve reactive power sharing and distribution according to capacity
ratio．

In order to solve the problems of complexity and uncertainty in the evaluation process of air combat
decision-making，an evaluation index set for air combat decision-making is constructed and a closed-loop
evaluation method based on cloud centroid method and consistency weighting is proposed． Firstly，the basic data
are normalized，and the corresponding air combat decision cloud model is established． Then，according to the
similarity and consistency of the indicators，the weight of the indicator set is constructed． Finally，the
appropriate threshold is selected to judge the validity of the evaluation and complete the closed-loop feedback of
the evaluation scheme． The simulation example shows that the scheme ensures the accuracy and validity of the
evaluation，avoids the stochastic problem of traditional methods，and truly reflects the decision-making quality in
air combat．

Because the sEMG( Surface Electromyography) is very sensitive to muscle fatigue，different patients
and electrode displacement，it is an arduous task to design a reliable robust and intelligent hand rehabilitation
device． To address these difficulties，a neural decoding method of rehabilitation gestures based on deep learning
is presented by using sEMG on the forearm of patients and CNN ( Convolutional Neural Network) to recognize the
movement intention． A combined feature extraction method is proposed to extract the combined features of each
channel of 8-channel sEMG． The combined feature includes 32 features which are wavelet packet decomposition
energy features，time-domain features and frequency-domain features． The eight channel features are formed into
an 8 × 32 numerical matrix and grayscale processed into a feature map，to train the convolutional neural network．
For five different gestures recognition，the classifier’s accuracy reached 98． 1%． Finally，according to the
classification results，STM32 I /O port outputs the corresponding PWM ( Pulse Width Modulation) signal，which
shows the feasibility of this method and laying a foundation for further control of rehabilitation glove movement．

Due to the limitation of on-board processing capacity，Beidou satellite uses the distributed algorithm to
realize inter-satellite link autonomous orbit determination． Since the distributed orbit determination algorithm can
only acquire sub-optimal solution，to improve the accuracy of Beidou satellite autonomous navigation algorithm，
the realization method of the centralized orbit determination algorithm for Beidou satellitesis studied． A
centralized orbit determination algorithm based on extended Kalman filter algorithm is proposed and the on-orbit
realization process of the algorithm is designed． Finally，the algorithm accuracy and engineering feasibility are
evaluated by using the LONGXIN 1E300 processor on Beidou satellites． The simulation results show that the
accuracy of the whole network centralized algorithm is better than that of the distributed navigation algorithm． The
simulation on the LONGXIN 1E300 processor shows that the centralized navigation algorithm has met the
requirements for satellite application．

In order to solve the human intervention problem of existing human target recognition and tracking
system，an automatic human detection and tracking system is designed，which is composed of embedded system，
wireless communication technology and upper computer． The automatic detection and tracking system is divided
into two parts: the upper computer and the lower computer． Using STM32F103ＲCT6 as the control unit，the
lower computer detects the position of the human body through the SHＲAP-GP2Y0A21YK0F infrared ranging
sensor，and then controls the steering gear． The steering gear is equipped with a camera to collect the video
signal which is transmitted to the upper computer using WIFI( Wireless Fidelity) wireless technology． The upper
computer is developed by Eclipse-Android system development platform，which can display the video signal in
real time in the monitoring center． It has been proved that the system is easy to install and operate，can
accurately realize the automatic recognition and tracking process of human body，and can be widely used in the
industry of indoor non-interference infrared work．

In order to meet the practical teaching needs of the geoscience instrument course，a transient
electromagnetic experiment teaching model system is designed． It includes four parts: a transmitting /receiving
integrated machine，a receiving sensor，an abnormal coil and a host computer control software． The transmitting /
receiving integrated machine uses the AＲM ( Advanced ＲISC Machine ) controller and CPLD ( Complex
Programmable Logic Device) as the main control unit，and drives the H-type full-bridge circuit to realize the
generation and transmission of a bipolar square wave． The high-sensitivity is designed using the differential
hollow coil and the amplifier circuit for the dual op amp sensor． By studying the relationship between the induced
electromotive force of the closed coil and the time constant，an abnormal coil is used to simulate a good
conductive finite conductor． The host computer control software is developed on the Visual Studio platform using
C # language． After testing，this model can collect and display the transient electromagnetic induction secondary
field attenuation signal curve in real time，realize the simulation of the working process of the transient
electromagnetic instrument in the laboratory，and have a certain auxiliary effect on the practical teaching of the
course．

In order to better promote the construction of new engineering and cultivate engineering science and
technology talents with cross-border integration ability，the multi-disciplinary integrated innovative analog circuit
experiment is proposed． The experimental contents of thick foundation，flexible module，multi-disciplinary，
thematic and innovative are designed． TBL-ＲBL( Team Based Learning-Ｒesearch Based Learning) two track
teaching method is applied． The knowledge of communication，signal and system and so on is integrated． This
paper expounds the design method of cross innovation experiment in analog circuit experiment teaching through
the design of visible light communication system based on ＲGB-LED( Ｒed Green Blue-Light Emitting Diode) ．
Interdisciplinary innovation experiment can broaden students’vision，cultivate the ability to solve complex
engineering problems and students’innovation ability．

Phase is important parameter for analyzing electrical signals． Usually the output of the phase
measurement circuit is based on absolute phase difference． This nonlinear input-output relationship makes it
difficult to realize high stability and low cost hardware circuit． A linear phase detection circuit is designed by
using operational amplifier and ＲS ( Ｒeset Set) flip-flop to realize the function of shaping and triggering the edge
of the double sinusoidal signal． Using operational amplifier LM6144，the mean deviation is less than 0． 3° when
the test signal frequency is less than 100 kHz． The design defect of nonlinear phase detection is solved．

In view of the problem that the traditional big data parallel mining method always mines all the data at
one time，resulting in a long mining time and low mining accuracy，the quantum computing is adopted to
optimize the incremental big data parallel mining method． Firstly，the parallel data mining model is built
according to the basic process of data mining． Then on the mining model respectively by defining a quantum bit，
quantum search algorithm，quantum neural network processing and mapping transformation，the incremental data
preprocessing，filtering weights are obtained by decomposition of matrix-vector multiplication，preprocessing
results by using the combination of parallel collaborative filtering． Finally，by quantum fuzzy clustering，large
incremental data parallel mining results are obtained． The experimental results show that the average recall rate
of the incremental big data parallel mining method using quantum computing is 97． 25%，the parallel mining
time is within the range of 2． 1 ～ 3． 2 s，and the accuracy rate is always above 95%． And this method has the
best convergence and strong optimization ability．

In order to solve the problem of energy hole and energy minimization in ring shaped WSNs ( Wireless
Sensor Networks) ，a FCＲA ( Fuzzy logic control based Clustering Ｒouting Algorithm) is proposed． Firstly，the
network is divided into rings with equal width，and the optimal cluster number of each ring is calculated with the
objective function for minimizing the energy consumption of each ring． Secondly，a fuzzy controller is designed to
elect the cluster heads． The input of the fuzzy controller is the residual energy of the node and the distance from
the node to the base station，the output is the probability that the node becomes the cluster head． Finally，the
inter-cluster multi-hop mode is used for data transmission． The weight function of each transmission path is
determined based on the residual energy of the node，the distance of the next hop and the number of hops to the
base station，and the optimal next hop relay node is obtained． The simulation results show that FCＲA can
effectively reduce the network energy consumption and prolong the network lifetime．

In the field of deep reinforcement learning， in order to further reduce the impact of value
overestimation on policy estimation in TD3 ( Twin Delayed Deep Deterministic Policy Gradients) and accelerate
the efficiency of model learning，a DD-TD3 ( Twin Delayed Deep Deterministic Policy Gradients with Dynamic
Delayed Policy Update) is proposed． The delay update step size of the actor network is guided by the dynamic
difference between the latest loss of the critic network and its exponential weighted moving average． Experimental
results show that compared to the original TD3 algorithm that obtain high reward value in the 2 000 steps，the
DD-TD3 method can learn the optimal control strategy in about 1 000 steps and obtain a higher reward value，
thereby the efficiency of finding the optimal strategy is improved．

In order to improve the accuracy of wind farm short-term power prediction，based on deep learning，
a method for integrating wind farm short-term power prediction using the Stacking algorithm to integrate multiple
GＲU( Gated Ｒecurrent Unit) models is proposed． This method first builds three multi-layer GＲU neural network
models establishing a first-level model，extracts high-dimensional temporal feature relationships in depth，builds a
training set from the prediction results of the first-level model，and then uses the newly generated training set to
train the second GＲU model． The second-level GＲU model uses a single-layer structure to find and correct
prediction errors in the first-level model and improve the overall prediction result． Finally，a two-level model
embedded Stacking fusion model is obtained． Taking the historical data of Ningxia Taiyangshan wind farm as an
example，the accuracy of the model is verified． The experimental results show that the GＲU model fused by the
Stacking algorithm has at least an increase in the average absolute percentage error index of 0． 63 compared to other
algorithms，and the overall prediction effect is ideal． The accuracy of prediction has been improved significantly．

In order to improve the performance of the traditional face recognition algorithm based on PCA
( Principal Component Analysis ) and SVM ( Support Vector Machines ) ，we introduce an improved crossvalidation
algorithm with grid search method to optimize SVM parameters，which combins with PCA and SVM
algorithm． The improved algorithm uses the K-CV ( K-fold Cross Validation ) algorithm to optimize SVM
parameters，minimize the impact of individual sample errors on the prediction model，shorten the search time and
improve the face recognition rate． Compared to other PCA( Principal Component Analysis) and SVM( Support
Vector Machine) salgorithms，this algorithm has 9． 08% higher performance than tradition algorithm．

In order to meet the requirements of ministerial level open fund project management，a comprehensive
analysis of the various requirements of project management is conducted，and an open fund project management
system is established based on a service-oriented architecture． This system is based on J2EE framework，B /S
( Browser /Server) architecture，Java language research and development，and the system is deployed in the
cloud computing platform． It can complete the function management of all links in the whole process of laboratory
fund project application，reviewing，mid-term assessment，settlement，and results sharing． The actual operation
shows that the system can run stably and efficiently in the cloud and meet the online concurrent management of
laboratory open fund projects．

With the continuous advancement of informatization and teaching management in colleges and
universities，the teaching management of colleges and universities is moving towards the direction of intelligence
and informatization． In view of the current innovative problems of intelligent campus informatization teaching
management models in colleges and universities，combined with the actual application the problems and
deficiencies are discussed in the causes of the problems，and the innovative path of the current smart campus
informatization teaching management model in colleges and universities is proposed to strengthen the advancement
of teaching management in colleges and universities and create favorable conditions for the modern development
of college education．

In order to solve the problem of multiplicative noise interference in MＲS signal，a parameter extraction
method based on Markov chain Monte Carlo is proposed． The priori information model and likelihood function
model of the complex envelope parameters from ground magnetic resonance ( MＲS: Magnetic Ｒesonance
Sounding) is established． A novel method based on MCMC( Markov Chain Monte Carlo) ，which sample and fit
the posterior distribution of the parameters is used to get the data with the most occurrence times of the posterior
distribution． The largest weight of the posterior distribution is used as the optimal estimation value of the
parameters． By comparing the extraction results of MCMC parameters under different noise conditions with the
nonlinear fitting method，it is proved that MCMC method can extract MＲS signal parameters with high accuracy
and stability，which is under the interference of multiplicative noise．

In order to minimize the huge economic losses to people’s living and the country caused by major
natural disasters and to establish an effective disaster response transportation system，we proposed modern
intelligent algorithms such as clustering model based on K-means and multi-person short-circuit model based on
genetic algorithm，and carried out disaster rescue simulation experiment combining with Puerto Ｒico’s urban
data，which improved the traditional low traffic disaster response system limitation，pertinence and low efficiency
of faults． And for hospitals，road network density，population density and plains the model will provide the first
aid． We use drones to quickly inspect the main traffic lines and quickly restore the traffic． Taking Puerto Ｒico as
an example，the experiment results show that the model quickly realizes the traffic recovery，provides great
convenience for the transportation of ground materials，and improves the rescue speed．