To make full use of existing spectrum, we study the problem of a joint overlay and underlay resources allocation for the OFDM(Orthogonal Frequency Division Multiplexing) based cognitive radio systems. A new subcarrier and optimal power allocation scheme is proposed under the joint overlay and underlay fashion. This scheme can provide maximum transmission capacity for secondary users while keeping the total power and the interference introduced to the subcarrier below a given threshold. In consideration of the computation complexity of the scheme, the nulling scheme can be used in the proposed method to obtain sub-optimal scheme. Finally,the analysis of the robustness of this power allocation algorithm is presented when considering the uncertainty of
the channel gain between secondary user and primary user in the sense of the Worst-Case. Simulation results show that the proposed schemes demonstrate better performance than the overlay or the underlay resource allocation schemes.

Aiming at effectively extracting the event information of seismic signal and solving the issues that the Hemite-Gauss steerable filters have an intrinsic limit which the same amount of smoothing is made in all directions due to the isotropic Gaussian factor, the EHGFs(Elongated Hemite-Gauss Functions) are proposed.This method considers that the desired signal is different from random noise in direction characteristic, generate EHGFs steerable filters by the approximate fitting, which can improve the linearity of the signal and filter along the event to get directional filtered data with less smoothing effect. Then we reconstruct the data on the basis of the digital local features to denoise. Experiment results indicate that this method can significantly suppress random noise and keep the useful signal in seismic processing.

Abstract: To complete the ECG signal feature points extraction and the classification of ST segment, we put forward an algorithm based on the discrete wavelet transform, combined with the f derivative and the SVM(Support Vector Machine). The algorithm can accomplish the signal preprocessing, noise elimination, QRS complex detection and extraction of characteristic value, calculating the average ST segment, curve area and the standard deviation, and the simple classification of ST segment by using the SVM combined with the three sets of data. The matlab simulation results show that the wavelet denoising is effective and has no distortion, and completely extract ST segment feature points. The data are downloaded from the MIT-BIT database, the classification results show that cross-validation average accuracy is 80. 70%,the average accuracy of training is 91. 83%, the average testing accuracy was 74. 28%.

Actual seismic exploration is corrupted by a variety of random noise. It is necessary to eliminate noise to extract the valid data. To address modeling method of ambient noise in land seismic exploration, this paper used classical chaotic Duffing system to model the ambient noise of forest belt area in China according to chaos of the ambient noise. It used MCMC(Markov Chain Monte Carlo) method to solve the system parameters according to nonlinear characteristics of Duffing system, and it contrasted simulated noise with actual ambient noise in terms of time-domain waveform, phase diagram, spectrum and Lyapunov exponent. The results show that the simulated noise records keep the chaos of the actual ambient noise, and the two noise have a good curve fitting in terms of time-domain waveform, phase diagram, spectrum. Therefore, the use of the Duffing system for land seismic ambient noise has certain prospects. The surrogate data obtained by the Duffing system is closer to the actual noise, which can provide new strategies to suppress noise and theoretical supports for seismic exploration in complex environment.

In order to achieve to detect a variety of physiological signals form ECG ( electrocarcliogram)monitoring signal, reducing the complexity of the monitoring equipment, according to the impact of breathing motion on the ECG, it presents a respiratory information extracted from the ECG (ECG-derived respiratory signal, EDR) algorithm. First, according to Pan & Tompkins algorithm extracted from the ECG R-wave, and then, using a natural cubic spline interpolation algorithm to estimate the baseline ECG and remove baseline wander, getting clean ECG, finally, get clean ECG R-wave, the use of natural spline interpolation algorithm is applied to obtain R wave amplitude modulation signal which contains respiration information. In paper, it selected Fantasia database, which in addition to providing the ECG signal, it also gave a breathing signal synchronized recording. It used Matlab software to verify the proposed algorithm, associated analyzed and compared. The simulation results showed that the algorithm mentioned in this paper can extract the respiratory
signal from ECG. By comparing with respiratory data in the database simultaneously recorded, it confirmed the validity of the algorithm.

To solve the increasing serious traffic congestion problem, we designed and implemented a set of real-time, convenient, low cost, high precision and low error ITS ( Intellcgent Transportation System).Using road set up cameras to collect real-time traffic image as the main object of study, under the embedded
Linux system, we use the OpenCV(Open-source Computer Vision library) to carry out the vehicle lane change behavior information statistical calculations, and transplant the complete application to the ARM11 microprocessor on which we realizes hardware implementation. The experimental results show that the system is accurate, stable and has a good real-time performance effectively avoiding the interference, has strong practicability.

To reduce the influence of spectrum leakage of spectrum analysis, the method and operation of traditional DFT(Discrete Fourier Transform) are expanded from classical one-dimension spectrum to a time-frequency spectrum of two-dimension, and the relationship between a frequency component's amplitude in energy leakage spectrum and truncating length of a time-domain signal is examined in terms of time-spectrum. The result indicates that the amplitude fluctuation of any frequency component along with the truncating length of the time-domain signal complies with sinc function, so that a expression of signal can be presented in time-spectrum with a sinc function. The spectrum from any appointed signal length is constructed and zero error of amplitude spectrum is implemented. The index law that the relationship between DFT spectrum amplitude error under un-period truncation and signal cycles is obtained. When the signal cycle length is 10 times, DFT spectrum produced by the standard deviation is about 0. 001.

In order to solve finite time rendezvous control problem for multi-agent networks where the dynamics of agents is modeled as a first order integrator, the control scheme is proposed based on the potential energy function. Assuming that only a subset of agents know knowledge of rendezvous objective, distributed nonlinear non-smooth control protocol is proposed based on state information and the potential function method, inspired by the principles of ther-modynamics. The finite time stability analysis is made according to the definitions of differential inclusion and generalized gradient and set-valued Lie derivative and invariance principle of non-smooth analysis so as to obtain the sufficient conditions that network can achieve finite time rendezvous.Finally, simulations are used to illustrate the effectiveness of theoretical results.

In order to maintain the higher reliability and desired dynamic performance of the system in the case of actuator failures, the problems of robust fault-tolerant control for a class of nonlinear NCSs (Networked Control Systems) based on T-S fuzzy model with uncertainties and time-delay are investigated. In consideration with the imperfect networked environment with quantization error, a global fuzzy model of the closed-loop system is established by using PDC (Parallel Distributed Compensation) algorithm. By constructing a delay-dependent Lyapunov function, utilizing Jensen inequality and introducing the free-weighting matrices, the conditions for the existence of robust fault-tolerant H_∞controllers of the closed-loop system are obtained, and the design of controllers is converted into a convex optimization problem of solving a set of linear matrix nequalities. A simulation example illustrates the effectiveness of the proposed approach.

In order to improve the real-time performance of predictive controller for vehicle yaw stability control,we implement a predictive controller based on FPGA(Field Programmable Gate Array), and adopt PSO(Particle Swarm Optimization) combined with penalty function to solve the QP(Quadratic Programming) problem. The FPGA is used to reduce operation time of the MPC (Model Predictive Control) controller by using parallel computations. In order to verify the effectiveness of the controller, real-time tests are carried out in the typical vehicle running condition. In the tests, vehicle dynamic model in veDYNA is used as the plant, and the FPGA is used as the hardware platform of controller. The simulation results indicate that the MPC controller can satisfy the requirements of the vehicle yaw stability control well, and lay the foundation for the controller to real vehicle test.

To study the changes of nonlinear dynamics characteristics with different rudder, height and mass, the continuation algorithm is used to identify the equilibrium surfaces versus elevator, height and mass. According to the equilibrium, the global stability of hypersonic vehicles is studied based on bifurcation theory. Compared to conventional linear method, the analysis results show that bifurcation theory can describe the dynamic characteristics more accurately, which can provide a powerful tool for the design of aerodynamic layout and control laws of hypersonic vehicle.

As crowd-sensing networks can efficiently address the key problem of large-scale sensor networks,i. e. large quantities of deployment and maintenance cost, it becomes a research hotspot for the Internet of Things. This paper presents the state-of-the-art of crowd-sensing networks. We first introduce the definition, the origin and the framework of crowd-sensing networks, followed by the analysis of its characteristics and the comparison with traditional sensing networks. And then, recent researches of crowd-sensing networks are surveyed in terms of environment sensing, public infrastructure sensing and social sensing. Finally, we point out three important directions for the future research of Crowd-Sensing Networks and analyze their challenges, to give a reference and help for the future study.

The current clinical measurement of BP(Blood Pressure) is mostly based on the intermittent type, and cannot get rid of the shackles of the inflatable cuff. On the basis of measuring blood pressure by pulse wave characteristic parameter method, a measurement based on SVM(Support Vector Machine) regression is proposed to realize continuous blood pressure measurement. Taken the pulse wave for the main object of study and extract their feature points, divided the pulse wave according to these feature points, then the time domain characteristics of pulse wave were obtained. To find the relationship between arterial blood pressure and these characteristics, a regression model based on SVM was established. The experimental results show that this method can measure continuous blood pressure well, the average error of BP measurement are all less than 0. 67 kPa, and the standard error are all less than 1. 07 kPa, this can meet the standards of the AAMI(Association for the Advancement of Medical Instrumentation).

The key technology of electronic image stabilization is the motion vector estimation. FPS(Feature points) matching based on the ROI(Region of Interest) area is used to estimate motion vectors. Firstly, ROI of the input image was selected, then the FPS were detected by Harris. Secondly, the rule of SAD inside the
window was used to sparse the FPS and find matching point in the adjacent frame. Finally, distance criterion was applied to eliminate false match points. Results of frame by frame motion compensation show the effectiveness of the algorithm.

In order to meet the low-cost high-speed, high-accuracy ADC ( Analog-to-Digital Converter)requirements, this paper presents a low number to the number of ADC to achieve a high level approach. The method uses a single chip as a comparator ADC, the signal is superimposed on a disturbance signal, after several analog to digital conversion and accumulating a number of high-precision ADC to achieve results. Experimental results show that the designed circuit is reasonable and correct, to achieve a low cost ADC in high precision measurement.

In order to evaluate the comprehensive performance of OVERHAUSER magnetometer, and provide theoretical guidance for the development of proton magnetometer, the methods of sensitivity characterization of OVERHAUSER magnetometer in time domain and frequency domain are studied. Theoretical research proves that the two methods are essentially consistent. The effect of sample rate on sensitivity is analyzed. Study shows that the signal quality is the main impact on sensitivity. Experiments are designed to calibrate sensitivity of JPM-2 type proton magnetometer, and experimental results indicate that the sensitivity of JPM-2 type proton magnetometer is 0. 18 nT when characterized in time domain, and 0. 32 nT/ Hz@0. 1 Hz when characterized in frequency domain. We obtain the frequency domain distribution figure of measurement data for diurnal variation component and noise component.

We presents a new method for human action recognition to solve the problem of representing the invariant property of the moving target, which combining the local motion pattern and kernel principal component analysis. Firstly, using the local motion pattern descriptors to represent the human motion and then using the kernel principal component analysis algorithm to deal with the local motion pattern descriptors to form a new feature description. The experimentshows that human actionrecognition based on LMP-KPCA, with the other two classical methods (Cuboids+SVM and LMP+SR)by contrast, the recognition rateis improved obviously, the corresponding improved recognition rate is 1. 1% and 1%.

Aluminum surface coating quality in production enterprises generally inspected by human eyes.Because the human eye can easily fatigue, and detection is inefficient. In view of this kind of situation, we proposed the method of image inspection based on the particle analysis of mathematical morphology and fuzzy
kernel clustering. Firstly we use morphology particle analysis for the image coating surface of the aluminum, then se parallel structure characteristic of the mathematical morphology by introducing kernel function to fuzzy kernel lustering in the feature space, to achieve the purpose of quality inspection. The experimental results show that this method can achieve convergence rate 18. 4 s, classification accuracy rate can reach 91. 6%, the overall performance is better than the traditional clustering methods, and has strong robustness.

In order to separate the DataMatrix code image from complicated background, to improve the accuracy and speed of the decoding, we use the cross correlation matching algorithm to locate the code, then we use the “pyramid-layered algorithm" to reduce the calculation cost. The improved Hough transform is used to carry out the “L" boundary line of the image. The whole algorithm and the decoding process run on S3C2440 chip equipped with the Linux system. The result shows that with this algorithm, the decoding time of DataMatrix code can be reduced from 2. 3 s to 500 ms, and decoding accuracy can be improved from 85% to 100%, so it can meet the practical demands.

To monitor the daily living of the old people living alone, a study of their activity amount based on time series analysis and fuzzy pattern recognition is proposed, in which the daily activity of the old people living alone is used as the effective index to monitor their living state. The predictive model for the activity amount time series is established by using random time series analysis. Using the fuzzy pattern recognition method, the difference between the amount of monitoring activity and the amount of forecast activity is identified. When the gap is more than the normal range, the warning information will be reflected to the guardian of the elderly. The result shows the accuracy rate of abnormal behavior recognition was 96. 97%. The study provides an alternative to activity amount monitoring of the old people.

To lower the cost of generation and optimization for electric customers experience, an economic scheduling model for a micro-grid is proposed considering DSM(Demand Side Management) with one solar source, two diesel generators and one battery. Demand-side electric customers respond to the dynamic pricing mechanism and transfer the load to achieve the purpose of saving electricity costs. Connecting the inconvenience caused to the customer with the duration of shifting and using genetic algorithm to minimize inconvenience and electricity costs of customer. The simulation results show that the cost of generation is less with DSM compared to the case without DSM and there is savings for the customer.

In order to effectively reduce the data dimension and remove the information of video data and improve the efficiency of abnormal event detection, the video abnormal event detection method based on the distinction and correlation is proposed. The method extracts features of the video data by analyzing the temporal and spatial neighborhood information of data, removes redundant information in the feature set and improves the efficiency of abnormal event detection by analysis of the distinction and correlation of feature. The method is compared with the traditional method in the simulation. The experimental results show that the detection accuracy of video abnormal event detection method based on feature fusion is higher than other methods, the method can accurately locate the abnormal area in the scene.

Fractal structure features of China's stock market by EMD (Empirical Mode Decomposition) and (R/ S: Rescaled Range Analysis) analysis method is explored. It is analyzed that the CSI 300 index closing price logarithmic return rate by using empirical mode decomposition method, using R/ S analysis of fractal theory to research empirical mode function to reveal the fractal characters of Chinese stock market. the final result shows that Chinese stock market has obvious self-similarity, the yield of the decomposed sequences is biased random walk process.