The exsisting enumeration methods or the subsequent modifications can not adapt well to signals of different properties. A SVD based (Singular Value Decomposition based) enumeration method of Hankel matrix of the spatial correlations of sensor outputs is proposed for the ULA (Uniform Linear Array). Utilizing the general form of Hankel matrix of sensor array with the exclusion of the term of the unknown noise variance from the correlations of sensor outputs, the proposed method can estimate the number of sources whether or not source signals are independent and enumeration capability is up to half of sensor number. The simulation results show that distribution pattern of the PCEs (Probabilities of Correct Enumeration) is invariable and the proposed method has more steady enumeration performance and wider threshold scope by comparison with that based on the spatial smoothing scheme.

In order to accurately evaluate the hybridradio frequency/ free space optical RF/ FSO (Rradio Frequency/ Free Space Optics) relay system performance, the closed form analytical expression of end to end outage probability of hybrid RF/ FSO relay system is derived by using the Meijer's G function. The hybrid
RF/ FSO relay system uses the amplification and forward relay and the RF(Radio Frequency) dual antenna transmission technology. The fading of the RF channel obeys the Rayleigh distribution, and the fading of the FSO (Free Space Optics) channel obeys the M distribution. The Monte Carlo simulation verifies the accuracy of the derivation.

To solve the weak robustness of extended Kalman filtering to the permanent magnet synchronous motor without position sensor, a rotor speed and load rotor position angle estimator based on H∞ filtering algorithm is proposed in this work. The robustness against model uncertainty and non Gaussian noises are improved. While the external load torque is also estimated as a system parameter. This method overcomes the impact on load torque changes through utilizing estimated value. The amount of calculation is decreased by using reduced order model of PMSM(Permanent Magnet Synchronous Motor). Finally, the stronger robustness can be shown with simulation results compared with the extended Kalman filtering.

A new coupling analysis method is proposed to solve the problem of coordination in the flight control system of hypersonic vehicle, which is used to study the coupling interconnection of the attitude dynamics for hypersonic vehicles, and a coordination controller is designed. Firstly, the attitude dynamic characteristics of a hypersonic vehicle is studied. Secondly, according to the coupling characteristics, a novel coupling analysis method based on the attitude dynamical equation is proposed to describe the coupling relationships among the attitude variables. Then, combining with the dynamical matrix, a coordination controller is designed based on hierarchical sliding mode control method to deal with the strong coupling problem between the control surfaces.Finally, the asymptotic stability of the closed loop system is proved by using Lyapunov theory, and the simulation results are given to verify the effectiveness of the proposed method, and the coordination control of attitude is realized.

The solution to the problems of low resolution and obvious stitching line in fusion area after image mosaic is proposed, by introducing irregular overlapping region generating method based on piecewise linear approximation to reduce algorithm error and using adaptive algorithm based on distance ratio to implement the automatic matching of weights, which can make sure the algorithm get more accurate results without being restricted to the shape of overlapping region. The simulation showed the improved weighted average fusion algorithm achieved the purpose of removing the stitching gap and smoothing the fusion area with the same rapidity as original algorithm's and was able to obtain large field of view images with better fusion quality.

Traditional model reference adaptive theory MRAS(Model Reference Adaptive System)of induction motor speed sensorless drive has a shortcoming that the reference model is not accurate. This paper proposes a adaptive speed sensorless control AFO(Adaptive Full-Order Observer) that uses motor itself for the reference model, luenberger full order observer for the adjustable model. The syetem consists of a full order observer that yields the rotor fluxes of motor needs and also estimates the speed using an adaptive law. Based on the Lyapunov stability theory. The adaptive rotor speed is actually derived by using estimated stator currents error and estimated rotor flux. A fast convergent feedback matrix is designed based on the pole assignment method,to ensure the stable operation of the motor in a wide range. In view of the instability of the rotor speed estimation in low speed regenerative braking, by using the Routh stability criterion, a new feedback matrix is designed to guarantee the global stability of the speed estimation. The simulations results show its validity and effectiveness.

Aiming at the challenges, such as fewer labeled samples and expensive manual annotation in medical images, a network of liver CT(Computed Tomography) images segmentation model based on semi-supervised ladder is presented. First, the input data is reduced by super-pixel segmentation. Next, the patches are
extracted around the center of pixels, and the patches are used to train a semi-supervised model. Finally, the trained model is used to achieve liver segmentation. Experiment results show that a small number of labeled pictures are able to obtain similar results with supervised learning.

In order to improve the machining accuracy of cam grinding and solve the grinding accuracy problem of cam grinding system, the contour control strategy of cam grinding platform based on equivalent error method and B spline curve is put forward. By using B spline interpolation method, the motion command is given, and the lift data of the cam is processed by B spline inverse algorithm to obtain the control points of the generated sequence, to interpolate operation. Using the equivalent contour error as the object, the nonlinear equivalent error model of cam grinding system is established, such two axis tracking precision problems become equivalent error stabilization problems. And the control input value is calculated for compensation of contour error. In order to make the controller and the B spline curve design instruction compatible with Sylvester implicit method, B spline curve parameters are converted into algebraic form,
combined with the use of two methods to design the controller for high precision machining CNC cam grinding platform requirements. The proposed control strategy is demonstrated on the Sinulink simulation platform, It is proved that this method is feasible and effective, reduces the contour error and tracking error
of the system, and has good contour performance.

In the process of predicting, the remaining useful life of Lithium-ion batteries is based on particle filter algorithm. The fundamental particle filter algorithm has the problem of particle degeneration and it is difficult to ensure the accuracy of the remaining useful life prediction, so an improved unscented particle filter algorithm based on MCMC (Monte Carlo Markov Chain) is proposed. This algorithm overcomes the problem of particle degeneration by selecting the appropriate importance density function and resampling strategy, and improves the accuracy of the remaining useful life prediction. The simulation experiment shows that the improved particle filter algorithm can track the decline trend of battery capacity better and achieve higher precision than the fundamental particle filter algorithm,which can provide a new idea for predicting the remaining useful life of Lithium-ion batteries.

In order to solve the problem that the traditional monitoring mode can not accurately and real-time monitor the atmospheric environment data of various regions, an air environment monitoring system of Internet of things based on SCADA (Supervisory Control And Data Acquisition) system is designed. The system uses a
portable gas monitor as the slave computer to collect atmospheric pollutant concentrations, uses configuration software to build a central monitoring platform to collecting and analyzing thereceived data. The system realizes receiving data from the slavecomputer, realizes data storage with the server, and realizes real-time data monitoring and historical data inquiry. The system can monitor the changes of atmospheric environment of nodes in real time, and can also query the historical data of monitoring nodes and generate historical reports. It has the functions of collecting data, storing data and pre-warning of gas concentration. The system solves the problems that the traditional monitoring system can not refine the monitoring scope and the monitoring data can not be updated and released in real time, and it provides technical support for more convenient, real-time and efficient monitoring of the atmospheric environment.

Indoor quadrotor with the speed of conventional inertial measurement unit to collect speed data has great error, unable to point in the process of actual flight hovering. Aiming at this problem based on vision, a set of small indoor four rotor aircraft is built, with the application of cascade ratio PID (Proportion Integration
Differential) control algorithm. The experimental results show that it effectively solves the indoor or closed environment in the process of the four rotor in hovering flight level drift problem, making the four rotor aircraft flying in the indoor environment smoothly.

In order to complete task scheduling in grid environment better, and overcome the limitation of the needle current scheduling model, combined with the characteristics of grid task scheduling is a NP ( Non-deterministic Polynomial) problem, this paper proposes a novel grid task scheduling optimization model based on modified SFLA(Shuffled Frog Leaping Algorithm). First of all, grid task completion time is taken as optimization indicators to establish mathematical model, and then leapfrog algorithm is used to find the best scheme for grid task scheduling, and leapfrog algorithm is modified to solve defect and improve the algorithm speed, the grid performance analysis were validated by GridSim. The results show that modified shuffled frog leaping algorithm can complete task scheduling, shorten the task completion time, balance the load of resources, and the performance was obviously superior to the other algorithm.

In order to prevent physical illness caused by improper sitting position, a wireless wearable human body posture monitoring system is designed. The TI CC2540 processor handles the x-axis, y-axis, and z-axis data acquired by the LIS3DH sensor in real time and transfers it to the handset client APP(Application) via Bluetooth 4. 0 protocol for data processing and display to correct the posture of the human body purpose. And the system can be achieved with the organic combination of somatosensory games, to achieve the purpose of exercise.

In view of the low accuracy and slowness of traditional SIFT (Scale-Invarivant Feature Transform)algorithm of tool image matching in complex environment, a SIFT image matching algorithm based on edge information is proposed. The Canny operator is improved by adaptive smoothing and improved Otsu algorithm.
The edge information of the tool image is obtained. The edge is expanded and closed by morphological operation, and the edge of the edge is removed. The SIFT algorithm is used to extract the stable feature points to achieve the matching of the tool image. The experimental results show that the improved Canny operator can eliminate the influence of the pseudo edge on the edge detection and have strong adaptability. For the tool image with scale change, scaling, rotation and partial occlusion, the improved algorithm is superior to traditional SIFT algorithm in terms of matching accuracy and matching time. Which is better than the traditional SIFT algorithm.

In order to improve the accuracy and overall rate of information retrieval, on the basis of studying the relevant technologies about ontology constructing and semantic retrieval, ontology and semantic are applied to education resources retrieval field. Based on the ontology of user input query keywords query expansion, the use of Lucene for the expansion of keywords search, an education resources semantic retrieval system is designed and implemented. The comprehensive recall rate of the retrieval model is 81.5% , and the comprehensive precision is 83.1% . The two indexes are all superior to the traditional
keyword seardh methods.

Cloud computing technology in reducing enterprise information infrastructure investment, improve the information system availability and fault tolerance at the same time, there are difficulties to build the platform,operation and maintenance costs increased, the complexity of computing platform is simplified to build the cloud,based on the cloud computing infrastructure for research, a comprehensive analysis of OpenStack system,combined with the network structure, business integration and server resource planning to explore the feasibility and technical architecture of cloud architecture, and then puts forward the general cloud infrastructure management platform construction method, simplifies the complexity of cloud computing platform, and on the cloud computing platform of automatic fault event trigger and event tracking process, enhance the complexity of the operation and maintenance of the cloud computing platform.

In order to solve the problem that there are obvious errors in the conventional models and the expected positioning cannot be realized, a kind of indoor positioning system is designed. The system adopts the standardized rssi ranging model. The experiment shows that the system can solve the problem of positioning error, the error is reduced by 15%, and the reliable result is obtained.