Computer Science and Technology

The following sections are included:

• Preface
• Committee
• Keynote Speech
• Contents

An embedded system under the development of the microprocessor is widely applied in the world. The basis of further research and development and application is the question of how to build a small, real-time, good performance of embedded operating system. This research paper describes the methods and process of latest u-boot-2015.07, latest stable version of the linux-4.2.1 kernel, yaffs2 file system and the root file system transplant to the target board based on S3C2440 processor. The results prove that post-transplant Linux runs well on the target board. Embedded system migration not only familiarizes us with in-depth Bootloader, Linux kernel, file system and other software, but also get to know the various hardware resources of the target board.

Irrigation and water conservation are an integral aspect of rural infrastructural construction but canal water distribution has a problem of low commonality in most areas of China. For example, the irrigation canal system water distribution was optimized through improved ant colony algorithm, based on the geographic information and remote sensing measurement and control technology in Da an district in Jilin province. People eventually started saving water and its production increased. In irrigation process of northern China, optimization of water distribution is of great importance in the field of agriculture. The improved ant colony algorithm can minimize the water loss of canal water distribution pattern in some constraint conditions to make the limited water resources more useful. The results are more close to reality where it is very worthy of popularizing the improved the ant colony algorithm.

Our purpose was to improve anti-ship missile warfare simulation models’ efficiency. We compartmentalized model component architecture according to anti-ship missile warfare logic. The Paper researched information on the interactivity and design way of model bean. According to JavaEE/EJB architecture, anti-ship missile warfare simulation system was designed as data layer, operation layer and application layer. Hence, we realized simulation system this way. The method improves effective reusability and sharing of anti-ship missile warfare simulation models. It can be applied in all kinds of weapon warfare simulation.

Out-of-band concealment is an effective way of concealing data, but is it feasible on Android? In the paper, we will firstly study the out-of-band concealment and VFS(virtual file system), learn how the file system works, and further figure out whether out-of-band concealment can be implemented on the Android system. After analyzing the process and operations what the file system does when it truncates a file, by a series of experiments, we finally realize the out-of-band concealment on Android system. To improve survivability, we should combine out-of-band concealment with other techniques.

Petri net is an effective tool to analyze the operating rules of the system. Petri net process can reflect the feature of the system fault propagation. Construction and definition method of Petri net fault detection process is presented and the fault performance features in the Petri net process are analyzed. Based on the above research foundation, a method of fault detection and location with Petri net process is presented. Firstly, Petri net model with input and output for fault detection is given. Then fault detection and location method with this model is described in detail. Finally, an example of fault detection in sequential circuit is given to verify the correctness of the method.

This paper discusses the methodology used in designing and implementing a full-system cc-NUMA multicore architectural simulator based on the open source gem5 simulation framework. We analyzed the performance and power consumption of multithreaded applications in cc-NUMA multicore processors and the experimental results of the PARSEC benchmark suite demonstrated that fine-tuning cc-NUMA multicore architectural parameters such as shared cache geometry and replacement policies can lead to significant performance improvement and energy reduction of multithreaded applications.

The Petri net method is a reliable and accurate way of analyzing protocol, albeit with limitations in resolving conflicts between protocol performance evaluation and the protocol validation. A methodology of using original Petri net as a model of protocol verification is proposed to solve this problem. Without changing any part of the structure of the original Petri net, the method introduces protocol performance evaluation by attaching time delay to transition, hence avoiding repeat modeling. Therefore, the model can evaluate the performance of the protocol and validate it at the same time. As an example, the paper validates 0-1 stop and wait for protocol in detail and then evaluates the performance of the protocol by translating original Petri net of 0-1 stop and wait for protocol into timed Petri net.

This paper proposes the helper thread prefetch request breakdown to differentiate between different performance gains achieved by different helper thread prefetch parameter combinations. Cycle-accurate simulation results of selected benchmarks in Olden and CPU2006 show that, the selection of appropriate helper thread prefetching parameters can maximize the potential of helper thread prefetching, based on the different memory access patterns of target applications.

With the continuous development of smart grid technology, the smart grid is moving toward the direction of ubiquitous integration, thus the heterogeneous network elements have different security access techniques. The core of the smart grid information Safe interactive Strategy switch lies in network terminal information awareness, the underlying data of the cooperation decision support. This paper proposes a novel security cooperation framework based on network collaborative, to design a security policy switch decision method. The vertical cooperation security management module can effectively reduce the time delay and energy consumption of information gathering phase of the framework, providing powerful data support for the switching decision method.

A multi-functional crop intelligent monitoring system, including internet and Arduino micro-controllers, function to upload humiture, soil moisture, light intensity, PM2.5, harmful gas and other real time data. The system's other functions include automatic watering of crops when the humidity data exceeds warning levels, broadcasting the condition of crops by Microblog, remotely controlling the watering and lighting by web and WeChat and remote video monitoring. Data collection of sensor data is conducted by Arduino UNO micro-controller and the data is uploaded real time into the internet by HLK-RM04 WiFi module. Software functions of the system include real time charts by Websocket, JavaScript in PC and mobile internet, and concurrently, the implementation of WeChat control and Microblog broadcast through Tencent micro channel public platform and Sina Microblog.

Memory allocator is an essential component of a program and it highly determines the overall performance of a program. The current general memory allocator falls short on performance because it is unable to determine the memory allocation information of the upper applications. This paper introduces an optimized mechanism that leverages compiler instrumentation to gather and analyze the upper-level memory allocation patterns. Experiments has shown that our optimized method could achieve better performance (10% increase in performance on average and 18% increase under intense conditions) while incurring low amounts of overhead.

A complete simulation platform for master and slave clock synchronization based on IEEE 1588 protocol is presented. Comparisons with the published results from hardware platform have shown that the simulation platform gives a good indication on the level of synchronization performance. It can also be used for choosing the right parameters from both actual hardware devices and IEEE 1588 protocol.

The conventional range-free positioning DV-Hop algorithm used in wireless sensor networks experience low positioning accuracy due to several factors. An improved algorithm: HSDV-Hop (Harmony Search DV-Hop) was proposed in this paper which provides high convergence speed and outstanding performance in solving global optimization problems as compared to the original DV-Hop algorithm. The improved algorithm establishes the optimization functions by the information of distance between the nodes and the location of beacon nodes, so the locations of unknown nodes are estimated. The results show that the proposed method can significantly improve positioning accuracy as compared to the original DV-Hop algorithm without increasing the sensor nodes hardware.

Most DNS system apply cache mechanism to improve the efficiency of domain name resolving. Setting up the TTL value of the cache item is the key issue of DNS performance optimization. This paper proposes a dynamic adjustment algorithm for DNS cache item. By using DTTL model, our algorithm dynamically changes the TTL value as the varying of resolving rate. With comparisons to modifying the zone file directly, our algorithm makes a tradeoff between the consistence and hit rate.

In a P2P (Peer-to-Peer) VoD (video-on-Demand) streaming system, some nodes may receive too many requests, which leads to overload. On the other hand, some other nodes may receive too few requests, which leads to low utilization. However, existing related studies cannot handle this problem effectively, because they don’t have an efficient dynamic load information management mechanism, and they don’t distinguish the difference of requests when transfer the nodes’ load. In this paper, to manage the dynamic load information efficiently, we design a load information management table for each node. Based on the load information, we propose a load balancing scheme which uses a request migration algorithm (LBRM). Through simulations, our scheme can handle the load imbalance problem effectively and improve the users’ playback fluency.

Peer-to-Peer Video-on-Demand (P2P VoD) over Internet has become an effective and popular way to provide video information to users. The knowledge of video popularity is very important for system operation, such as video caching on peer-to-peer network. The video popularity distribution at a given time is quite well understood. However, in this paper, we study how the video popularity changes with time, classifying them into different types, and apply the results to design a novel popularity prediction model, which is derived by analyzing the video’s user rating information as well as time-variation. Based on this model, we propose a dynamic popularity-based cache replication strategy (DPCR). The simulation results demonstrate that our proposed algorithm can greatly reduce servers' workload and improve the streaming quality.

Increasing network traffic requires higher performance in the processing ability of network devices. The routing table lookup based on single thread is not appropriate for usage in general-purpose multi-core processors to achieve parallel processing. Current parallelization of routing table lookup in application layer causes high overhead in data copy, network stack and conflict of shared memory access. Network Processor Development Kit (NPDK) is a development environment for the network processing platform. In this paper, we proposed a routing table lookup mechanism named Multicopy Parallel Processing (MCPP) provided by NPDK for the purpose of multi-core parallel processing. Based on multi-core parallel processing architecture of NPDK, the overhead of packet copy is eliminated and multi-copy technology efficiently avoids routing table access conflict. Moreover, in order to reduce the number of memory accesses, a lookup algorithm named Configurable Two-Level Operation (CTLO) was implemented. Lastly, we compare the forwarding performance between the MCPP mechanism and the traditional mechanism in NPE platform, which is based on the general-purpose multi-core processor FT1000A. The evaluation results show that the packet forwarding performance based on the MCPP mechanism outperforms that based on the traditional mechanism.

Mind wandering is a state transition from task-related thinking to task-unrelated thinking and it has possible negative effects on learners’ performance in online education. The detection of mind wandering can improve learners’ performance by signaling, intervening and restoring attention to the process of learning. Traditional methods of detection of mind wandering involve specific devices such as eye gaze or physiologic sensors, tracking movements of eye ball or detecting electro-physiological activities of neurons or skin. This paper presents a new detection method using the camera of a mobile device to detect mind wandering when learners are tasked to watch a teaching video. Through the constant image capturing of the learner’s face, his or her eye features can be captured and used to predict the learner's state of mind wandering if any, using methods of computer imagery and supervised machine learning. Through our experiment, we have attained an accuracy of 53%, which exceeded expectations, considering the hardware platform. The performance of the detection system based on the MCPP mechanism outperforms the detection system based on the traditional mechanism.

The rapid deployment of LTE networks in recent years has lead to the issue of LTE networks being unable to satisfy user requirements with the available bandwidth. This is also compounded by the appearance of various new services and the increase of client numbers. Wi-Fi offloading is the popular technology to ease LTE network congestion. In this paper, we propose the User Satisfaction Oriented Data Offloading (USODO), which is capable of offloading certain users' traffic from LTE networks to Wi-Fi networks according to user duration in Wi-Fi networks and the scheduling result. Through simulations, we can confirm that USODO decreases LTE network traffic and improves user experiences in networking.

Improving network transmission efficiency is an important issue in 10 Gigabit Ethernet (10GigE) transmission. FPGA functions by receiving, sorting, combining and repacking required UDP protocol frames, before sending them to the downstream receiver. This design has high efficiency and low delay through FPGA hardcore programming. The start, stop and necessary parameters for the frame reconstruction are controlled and configured by the upper computer software, providing easy modification and good adaptation. The design can effectively improve the utilization of 10GigE transmission link, optimize the network environment, and enhance the processing efficiency of downstream receiving device.

With the popularization of computer network technology, people are becoming increasingly dependent on computer networks, thus the reliability problem of computer network systems has become an important and pressing issue. This paper has identified and analyzed four influential factors of reliability and discusses the four effective methods to improve the reliability of computer network based on the connotation of computer network reliability.

The traditional search engines show incompatibility and inefficiency in the search in the Internet of Things (IoT). To improve the effectiveness and efficiency of IoT search, we constructed a new search framework for IoT that supports searching for entities and the entity information featured by multidimensional dynamic data and status. We proposed and explained the MAMS (Matching Analytical Mechanism in Searching) system, in addition to the similarity matching method based on multidimensional sensory data to fulfil the search framework. Furthermore, we designed a set of reasonable experimental environment with real world data gathered by the Smart Lab to verify the validity and efficiency of the framework and algorithm. The experimental results demonstrates the efficiency and suitability of the search framework, MAMS and the similarity matching method for IoT search.

Software Defined Network (SDN) decouples the control plane from the data plane, yielding a vast flexibility for networks. This paper focuses on the optimal placement for heterogeneous controllers. First, SDN networks are modeled using graph theory. Therefore, the controller placement problem is transformed into a specified graph partitioning problem. Then, an optimal controller placement scheme is proposed based on multi-level graph partitioning. Finally, Simulation experiments are conducted on real network topologies. The results indicate that our scheme can achieve nearly optimal controller load balance and is of application value to wide-area SDN deployments.

Data transmission is affected by factors such as multipath effect, noise, time varying and Doppler effect in HF channel. In comparison with traditional equalization, turbo equalization is a better approach to these problems. In this paper, we build a simulation model of turbo equalization in HF communication system based on the characteristics of HF channel, comparatively analyze of the effect of some parameters such as length of interleaver, number of iterations and convolutional encoder. Subsequently the burst waveform in MIL-STD-188-141b and the MMSE-based turbo equalizer was developed, thus providing a reliable basis for the application of Turbo equalization in HF channel. The simulation illustrates that the MMSE-based Turbo equalization can reduce the gap between the simulation BER curve and theoretical curve, alleviating the impact of multipath fading, noise and Doppler effect; providing a high value and extensive application prospect.

Using system break thought as a breakthrough point, the paper clears the connotation of equipment support resource network invulnerability. It discusses the damage model of equipment support resource network from the two sides of attack target and attack strategy. Based on the thought of complex network, it builds the indexes and evaluation arithmetic of equipment support resource network invulnerability, analyzes with samples and verifies the validity of the method, so it can provide technology approach and method for strengthening battlefield construction and equipment support resource defense.

The hybrid access mode is considered as the most promising access control mechanism for femtocell deployment to enhance mobile service quality and system capacity for cellular networks. However, it is a challenge to encourage femtocell base stations (FBSs) to adopt hybrid access and provide services to roaming unregistered macrocell user equipments (MUEs) in two-tier Macro-Femto networks. In this paper, we propose a novel incentive framework which the wireless service provider (WSP) provides economic compensation to motivate selfish femtocell holders (FHs) to open their resources for MUEs. A Stackelberg game is formulated to maximize the utility functions of both WSP and FHs. The unique Nash Equilibrium for the proposed game is characterized and optimal strategies for both tiers are developed. Simulation results show the utility functions of both WSP and FHs can be improved significantly, which develops the hybrid access mechanism.

The use of cloud computing offers tremendous value to IT business, due to its low cost, agility and rapid elasticity. The commercial public cloud infrastructure can be accessed over the Internet, which offers advantages such as high Return On Investment (ROI) and more efficient purchasing. However it also raises concerns about lack of visibility, security, reliability and regulations. The industry requires cloud service providers to continue providing stronger security and data privacy than ever before. In this paper, we discuss recent security concerns on the cloud from industry perspectives.

The paper proposes a trusted network connection frame based on the separation of controlling flow and data flow, giving priority to maintaining data stream transmission. This is done through controlling the flow of communication on both sides of the identity authentication and trusted state assessment, and updating the access control policies of both sides, so as to maintain data stream transmission control. On the one hand, it can guarantee the real-time transmission of business data and on the other hand it can improve the efficiency of communication participants’ trusted state assessment, which is suitable for the intensive real-time requirements of industrial and business data.

With the increasing widespread of network applications, the role of network security is becoming more and more important in computer networks. The analysis and discrimination of network data stream and then the discrimination of intrusion behaviors is an important direction of network security research. In network intrusion detection, this paper introduces a nonlinear regression method -- partial least squares to predict the network behaviors. At the same time, in the calculation of partial least squares residuals, the paper adopts the Kullback Leibler - divergence as an iterative calculation standard so as to improve the detection speed and accuracy.

With rapid development of modern computer network technology and the wide application of electronic business and electronic government services, encryption technology is an essential component of information security and research efforts on encryption algorithm has gained importance. This paper introduces the origins and principle of MD5 algorithm and analyzes the collision issue in MD5 algorithm and the impact of security breaches in the field of information security. Thus the detailed methods to improve MD5 algorithm application security are subsequently proposed.

With the rapid development of information technology, the Internet has become an important communications tool for information sharing. However this inevitably brings about a security risk and thus network security has gained significance in current information technology research. Therefore, an evaluation system of network security indicators is now an essential technique. This thesis focuses on the specific issues of network security evaluation using the ABC method and subsequent analysis of each influencing factor of campus network security evaluation according to the effects index is conducted according to the size of weights and networking safety evaluation of the fuzzy comprehensive evaluation method.

Certificateless public key cryptography is an attractive paradigm which combines advantages of both traditional public key cryptography and identity-based cryptography because it avoids using certificates and does not suffer from key escrow. In this paper, the author proposed a certificateless signature (CLS) scheme built upon bilinear pairings and proved its security in the standard model. The schemeturns out to be more efficient than other proposed CLS schemes in the standard model, and the signin algorithm needs no pairing operation while the reverse operation requires only three pairing computations.

To tackle the problems of encrypted cloud storage in a limited bandwidth environment, we proposed a scheme that divided encryption block of plaintext according to the content changes. The scheme based on the optimized AES algorithm and exhibits improved efficiency and security of data synchronizing in the limited bandwidth environment.

In this paper we propose a new approach to reduce excessive duplicate alerts and high false positive rates in IDS. We used an improved quantum-behaved particle swarm optimization (IQPSO) algorithm byintroducing multiple segment processing and absorption wall. During Alert aggregation processing, wecalculate similarity of alert by fuzzy membership function firstly, then optimize attribute weights a similarity threshold by IQPSO. Experimental results show that the algorithm is effective in alert aggregation and gives better results in reducing false positive rate and duplicate alerts.

Communicating finite state machines can be used to describe communication protocols; however due to cryptographic operations in cryptography protocols, they undergo state explosion when describing cryptography protocols. In order to overcome this problem, a modeling method of cryptography protocols based on modified finite state machine is proposed in this paper. Firstly, the finite state machine is modified by adding elements into the formal definition of traditional finite state machines to express cryptographic operations and the related communication entity of each operation. In addition, theformal definition of Finite State Machine of Cryptography Protocols (CPFSM) is provided. Secondly, the transition diagram of CPFSM is enriched by changing colors of the nodes in traditional transition diagrams, and adding stripes and flags into the nodes. Thirdly, the general steps to build the CPFSM of a cryptography protocol are described. Lastly, the method is proven to be efficient and valid in describing cryptography protocols and provides increased precision in simulating cryptography protocols.

Numerous defects can occur in the process of manufacturing display panels. In particular, a defect called mura is one of the most difficult defects to detect using conventional image processing algorithms. It usually appears as relatively dark or relatively bright regions with very low contrast against the background when the panel is viewed by an imaging device. We have developed a fast mura defect detection method based on the conventional watershed algorithm. The flooding step of the watershed algorithm is carefully re-designed to detect mura defects that exist both inside and at the boundary of an image. The just noticeable difference (JND) technique is used to quantify the level of the mura defects. Experiments show that the proposed algorithm is efficient and provides rapid detection of low-contrast mura defects.

Cryptographic misuse is an increasingly common issue in real-world systems. In this paper, we collected and summarized 224 cryptography vulnerabilities in the CVE database over the previous five years and analyzed the implementation of cryptography systems in 131 common Android application packages (APKs), to understand why certain cryptographic misuses are more likely to appear in certain scenarios. We present a systematic analysis about the pertinence between certain cryptographic misuses and the different characteristics of three actual platforms (mobile, embedded and server). Thereafter, we propose several lightweight countermeasures to alleviate cryptographic misuse.

Due to the rise in popularity of Internet applications, security problems are becoming increasingly significant. The various types of network security problems include the complicated structure of computer viruses and attack information that are well concealed from the network security system, requiring the deployment of a large number of heterogeneous security devices search for every possible threat in order to maintain security protection. However, these isolated systems lack effective linkage and the large amount of information they produce is difficult to be organize and utilize, hampering the administrative staff's ability to monitor real status of network and to make timely response to attacks. This paper proposes the fusion system of network security incident management based on machine learning technology to extract these underlying security incidents. The improved expectation maximization algorithm was introduced to make preprocessing and correlation analysis for the original security data and to generate hierarchical standardized data step by step to identify the source of threats, and provide convenience for administrative staff to make effective responses.

Currently, the damage potential of Sybil attack is serious in ZigBee. In this paper, our proposed Sybil attack detection method allows the cluster head to identify different Sybil nodes by comparing the RSSI value and geographic information to make an accurate detection of Sybil attack. Both the security test and the experimental results demonstrated that the proposed method can detect Sybil attack efficiently at small expenses of communication and calculation.

A quantum key network that shields equipment and communication networks is proposed to address the issue of complicated key relations of equipment in Quantum Key Distribution (QKD) network. Taking into account the key relations of nodes in quantum key network, a quantum key network model to study the key relations in QKD network is developed using graph theory. The paper firstly provides a detailed introduction to the principle of QKD and its limitations, then studies the topology structure of QKD network and mechanism of the quantum key distributed service, and finally adopts graph theory to construct a key relations based quantum key network model to study key distributing processes in QKD network.

Cloud computing provides convenient, economic and highly efficient services. However, with its rapid development in recent years, cloud computing’s security issues are of increasing importance and data storage security has emerged as the most pressing issue currently. In this paper, we studied cloud computing and data storage security issues and proposed the MC-R strategy to tackle customer data privacy problems. MC prevents data leakage during data download and upload through masking and concealing data in the client while R enciphers data with the computing prowess of cloud computing. Lastly, we tested the strategy and through data analysis, the result of the tests demonstrates that it can successfully mask and conceal data and protect data privacy.

In this paper, we studied and proposed solutions to the hidden dangers in the security mechanism of Hadoop cloud computing cluster. By analyzing the security mechanism of Hadoop cluster, the paper identifies several security vulnerabilities as such user authentication and certification, and proposes the several solutions to improve the security mechanism of the Hadoop cloud computing. Regarding the data leak that users experience during download or upload, we present a digital envelope technology to encrypt the data transmitted. Additionally, a cryptographic algorithm is adopted to encrypt the virtual disk in cloud disk to prevent data lost. The improvement to the dispatcher in the cluster can enhance the stability of the host node and avoid the system dysfunction. Experimental results demonstrate the effectiveness of the proposed strategy.

A power information network risk assessment method of power information network information based on vulnerability analysis graph mode is proposed. The results of several vulnerability scanning tool are the data source of this method. The accuracy rate of one single scanning tool is considered as he reliability and the set of scanning results is considered as a reliability vector. By analyzing every reliability vector, the vulnerability reliability in a single vulnerability analysis graph can be calculated. A risk assessment is given in a small-sized power information network using the introduced method under three attack circumstances with varying periods of time. The experimental results demonstrate that the assessment of power information network risks is viable for this assessment.

This paper proposes a security linkage strategy algorithm to accomplish the performance evaluation of network safety, which effectively decreases the amount of computation and improves efficiency. Moreover, different security applications have their own requirements for safety, so that this paper thoroughly considers the security service types of various applications. The methods to compute network safety scores based on fuzzy logic are designed for four different kinds of services, fulfilling security applications’ requirements. In addition, this paper gives full considers the situation that current network conditions are not representative of future situations, but current network status changes usually show certain trends which can predict the future conditions, thus providing more comprehensive network information to security linkage strategy.

The attack-defense graph is a model-based network vulnerability analysis technique. Utilizing the characteristics of State Grid, this paper extends the attack-defense graph model and proposes a state attack-defense graph model. The state attack-defense graph use rules to model the attacker, and display all threat propagation paths which are generated by the attacker exploiting the dependence relations among vulnerabilities in the target network. In conjunction with prevention and control measures for vulnerability, the corresponding protection solution is proposed. In order to automatically generate the state attack-defense graph according to the network’s topology information, reachable relationship of nodes and vulnerability information, a generating algorithm of the state attack-defense graph is proposed.

The password-based authentication scheme requires users to use low-entropy passwords for identity authentication, thus the authentication process does not require public key infrastructure and users to store long symmetric key security hardware; providing high user convenience. Existing low cloud authentication security technology is not able to efficiently resolve highly complex problems, therefore the proposed password-based authentication method provides a simple solution to maximize the advantages of a password-based authentication scheme. We proposed several improvements in the scheme and analyzed its security capabilities to verify its feasibility for public cloud or private cloud environment.

In order to tackle the security problem of image transmission across public networks, a new image encryption algorithm based on unified chaotic system and lifting wavelet transform is proposed in this article. Firstly, the image is decomposed horizontally and vertically by lifting wavelet transform to obtain sub band images. Secondly, utilizing the sort permutation and the binary XOR operation from the unified chaotic series, the algorithm can achieve the encryption of the sub band images. Thus, the recipient of the images can correctly restore the sub band images and reconstruct the original images by a decryption algorithm. Finally, the analysis of algorithm security is conducted with theory analysis and computer simulation. The experiment results demonstrate that the new chaotic encryption algorithm based on unified chaotic system and lifting wavelet transform is feasible and provides higher security.

To tackle the security issues of massive network data, this paper utilizes the fuzzy Kalgorithm and Naive Bayesian classification to develop a joint classification algorithm, and Mahout technology to develop a classification algorithm for feature extraction based on the MapReduce framework of parallel computing. The establishment of the corresponding model and training allows the joint classifier to determine the abnormal traffic, so as to develop a comprehensive network anomaly traffic detection system, to improve the security of the massive network data processing platform.

With the increasing need of computer network intrusion detection for efficient, lightweight and real-time analysis, a real-time dendritic cell algorithm for intrusion detection is proposed to solve non real-time problems of classical system. The antigens are randomly presented by dendritic cells in time series. Once all antigen cells have migrated, the antigens will be immediately assessed and the output will be generated, the characteristic of real-time analysis. The pseudo code of real-time analysis dendritic cell algorithm is also presented, and it is tested in the KDD CUP99 dataset. The results of the experiments demonstrate that the real-time DCA has high detection accuracy and excellent real time performance.

This paper firstly provides a formal definition of the n out of n extended color visual cryptography scheme (ECVCS) based on the XOR cipher, where a color secret image and k color cover images are encoded into n shares in such a way that the n shares present cover images individually and any n shares and above will recover the secret image while any less than n shares, will not reveal any information about the secret image. The proposed scheme in the paper has no extra expansion and it can easily trade the visual quality of the recovery image with the visual quality of shares by changing the value of a certain parameter in the scheme. The validity of the proposed scheme is verified by formal proofs, and its feasibility is demonstrated by computer simulations.

This paper proposes a car key tooth code automatic identification method based on the spectral clustering method to tackle the problem of the current low efficiency and high costs of artificial and mechanical identification methods of the car key tooth code. By using the general auto key fixture, the tooth height of different keys is transformed into to an easily identifiable single detection chip height, which is associated with the contour of the car key image. The car key tooth code is obtainedby calculating the vertical coordinate of the extreme points of the contour line in the car key image, which are processed in accordance with the following method. Firstly, the key tooth edge is obtained with the spectral clustering classification method and the extreme points at the edge of the toothare determined with the minimum and maximum method. Secondly, the relative heights of the extreme points to the key's horizontal edge are determined and the corresponding key tooth code is obtained. Lastly the effectiveness of the method proposed in this paper is verified by a car key image.

Image retrieval technology has experienced rapid development in recent years and the image retrieval methods based on gist features and SIFT features are increasingly getting more popular. These image features are high-dimensional features and could potentially be simplified to improve the accuracy of image retrieval. In this paper, we studied the existing methods, moved the calculation of the distance between the image and image database offline and ranked the images offline. During a user search, only a few function values are required to be calculated to find the closest image. Based on the ideology of the algorithm, we have carried out the image retrieval based on gist and SIFT features respectively. The experimental results demonstrate that the speed and accuracy of the image retrieval areimproved when the image retrieval is applied to the gist features. The accuracy of the image retrieval is improved by 6.76% when image retrieval is applied to SIFT features. Therefore, the pre-ranking algorithm is an excellent image retrieval algorithm for dealing with high-dimensional features.

This paper proposes a custom E-R diagram to organize the MongoDB dataset schema to utilize its free model and dynamicexpansion characteristics. This allows for NoSQL data source integration of the heterogeneous data and the tackling of problems regarding heterogeneous data integration. Additionally, an algorithm is developed to convert global GSQL to local MongoDB query statement through the metadata of MongoDB.

Based on background knowledge of the "digital silk road" construction, this paper researches the digital speech technology which can support silk road of economic development of western minority regions. With regards to voice information processing, Mel-scale Frequency Cepstral Coefficients(MFCC) is the speech characteristic parameter frequently used. Human auditory sensing characteristics concerning the variation of feelings to different frequencies can be detected by this parameter. Therefore, MFCC is particularly suitable for use in speech synthesis. Based on the theoretical principle of MFCC parameter extraction, this paper conducts an in-depth research on the configuration parameters and HMM modeling and applies it into the parameter extraction of Tibetan speech synthesis. This lays a solid foundation to the realization of HMM Tibetan speech synthesis and provides technical reference for speech recognition and synthesis of minority languagesfor mobile clients.

The emerging web text data contain features such as obscure concepts and uncertain word boundaries. Given that the existing traditional index cannot support it efficiently, the paper proposes a distributed classification-based hybrid index model by combining the suffix array’s ability to support phrases with uncertain word boundaries and the reverse index’s high query efficiency and its ability to support sorting query results. It takes advantage of the Map/Reduce frame’s ability to presort web texts and develop a Map/Reduce frame-based distributed hybrid index model on its own classification results set.

This paper proposes a method to rapidly identify abnormalities in the big data of the seismic wave signal based on SURF algorithm with Hadoop cluster. The model matches the earthquake precursors waveform data images of earthquake monitor stations with the historical earthquake templates one by one, to rapidly determine whether similar changes in the past exist. The experimental results demonstrated that in comparisons to the conventional seismic data abnormalities detection method of wavelet transform, the improved SURF algorithm is 12% better in scales, and has a 13% improvement in matching results. The SURF feature matching algorithm based on Hadoop clusters meets the requirement of the storage and parallel processing of the massive earthquake data, and improves the abnormal identification efficiency of the big earthquake data.

Chinese abbreviations have been widely used in modern Chinese, and are one of the main sources of unknown words, resulting in the difficulty for correct Chinese character processing. This paper proposes a new approach called tr-HMM (TimeRelaxed Hidden Markov Model) to recover Chinese abbreviations to their root words. The tr-HMM is a transformation of HMM, which we have devised as the basic recovery strategy. It firstly considers the abbreviation as the output sequence, and the corresponding root word as the hidden sequence. Subsequently, the tr-HMM relaxes the time invariant hypothesis in HMM, and extends itself to a non-stationary HMM by making utilizing the time information. Lastly, we utilize the ten folds ten rounds method to prove that the tr-HMM is more effecttive than traditional recovery methods for the Chinese abbreviation recovery problem. The precision of the tr-HMM abbreviation recovery was demonstrated to be 84.3%.

With the rapid development of the e-business, social networking and enterprise information system, the volume of data produced by individuals and enterprises is increasing rapidly, resulting in a rise of cloud storage as an increasing number of enterprises and individuals store their data in the cloud for greater flexibility and economic savings. However the risk of leakage of sensitive user data is still prevalent. Although several keyword search schemes over encrypted data with privacy preserving has been previously proposed, we propose a novel scheme to support the multi-keywords fuzzy search while also supporting Chinese search. We utilize the locality-sensitive hashing technique to solve the fuzzy match and the experiments on real-world data demonstrate that our proposed scheme is accurate, efficient and secure.

As the energy consumption of storage systems is grows at a staggering rate, hybrid clusters have gained increasing importance as a potential approach to tackle this challenge. By introducing low-power nodes, data-driven companies like Facebook and Baidu have reduced the energy consumption effectively in their Master/Slave based storage systems. However, the Master/Slave based systems have several typical disadvantages such as low scalabilities and single points of failure. The P2P based systems with high scalabilities utilizes file location algorithms instead of table lookup mechanisms, thus resulting in a problem of how to utilize the different storage nodes discriminatively. In this paper, a hierarchical storage strategy called vnode hierarchical remapping (VHR) is proposed based on ’a P2P distributed system called ZDFS. The strategy guarantees the high scalability and viability of ZDFS, and takes advantage of different storage nodes. Several test cases running on X86 and ARM hybrid clusters are carried out, and the test results demonstrate that the VHR works well, it achieves a good I/O performance and low data access response time while reducing the energy consumption by 44.8%.

The index is one of the most important component in NoSQL database and the B-tree is the currently most widely used index structure. In order to improve the writing performance of B-tree, Log-Structured Merge-Tree (LSM-tree) has been developed. LSM-tree improves writing throughput, but its reading performance is hampered, especially for range queries. In this paper, we propose an indexed LSM-tree (iLSM-tree) to improve the reading performance of the LSM-tree by adding an extra index. To evaluate its performance, we implement iLSM-tree on HBase and perform a series of experiments to compare the iLSM-tree with original HBase which uses LSM-tree. The results demonstrate that the iLSM-tree has improved efficiency.

A feature word vector based on short text clustering algorithm is proposed in this paper to solve the poor clustering of short text caused by sparse feature and quick updates of short text. Firstly, the formula for feature word extraction based on word part-of-speech (POS) weighting is defined and used to extract a feature word as short text. Secondly, the word vector that represents the semantics of the feature word was obtained through training in large-scale corpus with the Continuous Skip-gram Model. Finally, Word Mover’s Distance (WMD) was used to calculate similarity of short texts for short text clustering in the hierarchical clustering algorithm. The evaluation of four testing datasets revealed that the proposed algorithm is significantly superior to traditional clustering algorithms, with a mean F value of 55.43% on average higher than the second best method.

Currently, there are no components identification and extraction methods are that further specify the components retrieval targets from the software design document. Therefore, this paper proposes a method for components identification and extraction based on XML document by using a multi-agent system to learn keywords and semantics from the XML document transferred from UML-based software design document. Thereafter OWL, a descriptive language, is used to verify the correctness and completeness of the derived results. The experiment verifies the effectiveness of the proposed method, thus providing a reliable foundation for future works on component retrieval.

Low quantities, structural complexity, prone to variable degradation and numerous failures, are some of the characteristics of an aero-engine and thus it is a great challenge to apply the fault diagnosis and preventive maintenance. In order to overcome such problems, we put propose the missing data processing model based on the expectation maximization algorithm. We estimate the parameters’ point estimates of Weibull proportional hazard model using this algorithm and calculate the maintenance interval based on aero-engine’ availability constraints.

Existing RDF keyword search studies focus on constructing smallest trees or subgraphs which contain all query keywords, but neglect the semantic association between RDF data. Thus, this paper proposes the keyword parallel search over RDF data based on semantic association (KPSRSA)) algorithm which utilizes a score function to measure semantic association by combining OWL ontology and the probability model. It uses a distributed database Hbase as a storage medium and Mapreduce to perform parallel query, which queries sub-clusters with semantic association in Map phase and constructs a series of associated clusters as query results in Reduce phase. The experimental results demonstrate that the KPSRSA algorithm improves the precision and relevance of search results and keywords. In addition, distributed storage and parallel computing inquiry has improved scalability.

Currently, information search on websites is an indispensable part of daily life. It is an important platform to obtain resources, such as regular notifications or current affairs about relevant companies, schools or other organizations. However, websites are prone to tampering by malicious attacks and thus tamper detection for websites is an important countermeasure to maintain the credibility and integrity of information found on websites. In this paper, we propose a highly feasible batch website tamper detection method based on text comparison. Different algorithms for text comparison have been proposed to analyze various websites with with varying degrees tamper detection using three factors: time efficiency, accuracy of detection and memory consumption. The experiments demonstrated that different algorithms have varying performances in terms of website tamper detection but nevertheless the string comparison method has more advantages under normal circumstances.

The genetic algorithm has strong capabilities in global optimization but it has a high likelihood of falling into local optimums. In contrast, the simulated annealing algorithm has strong capabilities in local optimization but lack capabilities in global optimization. By combining the advantages of simulated annealing algorithm and genetic algorithm this paper proposes a hybrid simulated-genetic algorithm (HSGA) to perform text clustering. We use HSGA specifically to optimize the initialization of the cluster centers for every iteration in the K-means clustering algorithm that is used to cluster text data. In addition, we also utilize the adaptive genetic algorithm to extract the eigenvalues of the text data, which performs dimensionality reduction and further improve the performance of the clustering. Experimental results demonstrate that the Hybrid simulated-genetic algorithm for text clustering not only improves the accuracy and recall rate of the text clustering, but also exhibits high efficiency.

Ontology is an effective method to solve the problem of heterogeneous atmospheric lidar data. The traditional method of ontology storage cannot be directly applied to atmospheric lidar data and thus this paper proposes an ontology storage model which establishes a mapping mechanism from the ontology of lidar data to the object-oriented data model (OODM). The atmospheric lidar data storage model is developed by analyzing the characteristics of atmospheric lidar data and the ontology syntax of OWL DL. The experimental results demonstrated that this storage model has high query efficiency and is viable for application in the sharing and usage of atmospheric lidar data.

This paper researches on the application of a user interest model in personalized search. For a given user, we view his query history and documents of interest so as to mine user interest data from him or her. According to the vector space model, we represent the user profile as a vector, which is formed through classification of a select group of user query words. The new approach used in this paper is that the vector of user interest is constructed from the user profile. Experimental results demonstrate that this approach exhibit good description performance with regards to user interest and it lays a reliable foundation for future research on personalized search.

The quantity of astronomical data has been growing at an exponential rate, creating unprecedented challenges in astronomical data mining. The rapid development of distributed cluster technologies and cloud computing platform provide new research ideas and methods for massive data processing and analysis. The distributed cluster technology Spark is experiencing a meteoric rise, and shows comparative advantages in terms of iterative machine learning and interactive data mining applications. In this paper, we use the latest release of Sloan Digital Sky Survey photometric data set to explore the suitability and application problems of the data mining technologies based on Spark in the massive astronomical survey data, providing a new methodology for large-scale astronomical data mining.

Cross language information retrieval(CLIR) is a research topic in the field of information retrieval, solve the problem of cross barrier of language is CLIR's core problem. This paper proposes a method with cross language information retrieval of Tibetan Chinese based on dynamic dictionary for query translation and translation ambiguity in cross language retrieval. The Tibetan Chinese electronic dictionary as based knowledge sources, combined with comparable corpus of Tibetan Chinese, extraction of Tibetan Chinese bilingual theme word pair, thus real-time updating knowledge source to solve the problem of the limit of entry number in knowledge dictionary source. According to the characteristics of cross language information retrieval of Tibetan Chinese, using translation expansion techniques and Translation Balanced method, which reduces the number of ambiguity translation, improves the accuracy of translation.

Checkpoint is a classic method for fault tolerance but its efficiency is hampered by a large overhead due to the dumping of large quantity of data. This paper proposse a new checkpoint method which allows for the monitoring and recording of all functions that modifies the data instead of directly copying the data. By replaying the corresponding functions, we can restore the data to the latest state. We demonstrate that this record and replay mechanism is perfectly suitable for the kernel data of a process. Experimental results have shown that our proposed method can effectively reduce the amount of checkpoint data and improve performance.

Maintaining a healthy diet is influenced more by the types of food you eat rather than the quantity of food that you eat. In this paper, a multi-objective healthy diet recommended algorithm is proposed according to individuals' dietary records and health conditions. Based on the multi-objective genetic algorithm (MOGA) and taking into consideration the recommended dietary standards and individuals' health conditions, the algorithm evaluates the cumulative and long-term effect of an individual's diet in order to generate a personalized healthy diet recommendation. Experimental results demonstrated that the proposed algorithm is more suitable for application in a personalized recommendation service than the algorithm based on random weighting algorithm.

To solve the inefficient bottleneck problems of the traditional relational databases in big data storage and access, this paper proposes an efficient approach to store big proteomic data based on NoSQL (Not only SQL) database. MongoDB, a typical software of NoSQL, is compared with MySQL in data storage and query performance. The experiment results demonstrate that the query and read speed of NoSQL database has been improved significantly as compared to the traditional relational database, especially for mass unstructured and semi-structured data.

Currently, rapidly locating relevant data sources and assembling the most appropriate data on the Internet is a tremendously difficult task for researchers. To address these issues, this paper proposes the development of a high availability ontology-based water environmental data retrieval and visualization system (OntoWE). The OntoWE prototypes have two significant features. Firstly, it establishes a water environment domain ontology to offer facilities to alleviate semantic heterogeneity and associate semantic information with the data retrieval process. Secondly, it embraces the capability of Service Oriented Architectures (SOA) and Silverlight to leverage the latest protocols of several open web service standards with the Managed Extensibility Framework (MEF). This system provides a centralized and easy-to-use interactive system which enables users to do a one-stop search, access and visualize different types of water data in a single environment. The feasibility and effectiveness of OntoWE system was demonstrated through several investigations about water quality data discovery and retrieval at basin scale in China.

Low dose X-ray CT reconstruction is currently a significant research topic in medical imaging. Thanks to the development of compressed sensing, iterative reconstruction with sparse regularization can obtain more satisfactory results than the filtered back projection. As classical total variation cannot reconstruct with noise polluted projection views, we propose a Gamma regulation in tomographic reconstruction, and the experimental results demonstrate better compatibility either in objective evaluation or visual sense than total variation.

This paper proposes a new fault diagnosis scheme based on continuous density Hidden Markov Model (HMM) for vibration signals. Features extracted from vibration signals of rotor-gear-bearing transmission system are used to train HMMs to represent various running conditions. The feature vectors based on the node energies of wavelet packet decomposition are extracted from the vibration signals. Faults can be identified by selecting the HMM with the highest probability. The proposed method was tested by measuring the data of rotor-gear-bearing transmission system and has been demonstrated to be accurate and feasible.

In this paper, we developed a RBM classification model and tested its performance. The RBM transformer fault diagnosis classification method increases the training of pretraining sets and provides an improved average rate of correctly classified faults. This method is applicable to a large number of training samples and has high scalability. Comparisons to the BPNN and SVM methods demonstrate that RBM achieves on average a higher accuracy rate of fault diagnosis, which is suitable for engineering needs.

To tackle the premature convergence problem of the Particle Swarm Optimization, we proposed an improved algorithm was proposed. Chaotic sequence is utilized to pretreat the positions of particles to introduce randomness, and then the reverse learning ability of the worst individual particle is introduced into the velocity update iteration of the particle. The simulation results demonstrate that the improved algorithm significantly decreases the premature convergence of particle swarm optimization.

This paper proposes a new feature for robust speech recognition. This method combines the benefits of differential powerspectrum and the power law. The extracted features from this combined method are known as mel frequency differential power cepstral coefficients (MFDPCC). The speech recognition performance of MFDPCC features is compared to the conventional mel frequency cepstrum coefficients (MFCC) features using fuzzy radial basis function (FRBF) neural network under different noise level and different vocabulary. The experiment results demonstrate that the proposed MFDPCC feature has higherrecognition rate and better robustness than MFCC.

Independent Component Analysis (ICA) is the focal point in blind signal processing (BSP). It aims to separate the relatively independent signals from the mixed signal source. FastICA is widely used because of its simplicity and fast convergence rate. However, it is extremely sensitive to the initial values, which affects the separation effect and even inconvergence if the initial values are not chosen appropriately. In order to solve the problem, a new method based on PBILto combine ICA is proposed. The learning rate and the learning model in PBIL are used to optimize the separation matrix in ICA, before the kurtosis is used to evaluate the effective values. The proposed method effectively avoids the problems in sensitivity of the initial values. Experiment results from the mixed speech signals data shows that the method has obvious advantages over FastICA on both separation effect and algorithm performance analysis.

Image segmentation is an essential research topic in the field of image processing, which primarily functions to separate figures into multiple disjoint regions. In most computer imaging scenarios, segmentation serves as the primary procedure for the further image understanding. In this paper, we propose a novel algorithm based on deep neural network and modified fuzzy clustering model. We enhance the clustering model with the incorporation of Markov random field theory tobuild up the spatial information model. To resolve the challenges influencing the performance of the traditional neural network, we adopt the deep neural structure to enhance the feasibility and robustness of the network. We compare our model with other state-of-the-art algorithms, using numerical and visual simulation to test the effectiveness,. Our experimental results verify that both accuracy and efficiency are promoted simultaneously. We attained a high segmentation accuracy of 93.6% with our proposed method.

A color cast correction algorithm based on improved Frankle-McCann Retinex is proposed to correct images which are influenced by illumination. To improve on the original algorithm, the distance-weighting factor with Gauss function is introduced, and a linear stretch with the mean and the standard deviation is carried out. Experimental results demonstrate that the algorithm in this paper has improved correction effect on the color cast image.

The difficulty of simulating Chinese ink animation in computer is a prevalent issue and in this paper, we propose theuse of Navier-Stokes equations to control the motion of the Chinese ink to form animations. This method can reflect the intrinsic physical properties of fluid motion, and it can realistically simulate various natural and physical phenomena. The equations were resolved by the semi-Lagrange method in this paper. The calculation of the equations is too complicated for real-time simulation, so GPU (Graphics Processing Unit) is used to accelerate the calculation rates. Experimentalresults have demonstrated the feasibility and viability of the proposed method and it greatly simplifies the production process of Chinese ink animation.

There are two major limitations with the existing retinal blood vessels segmentation method: high computational complexity and poor real-time capability. In order to overcome these limitations, this paper proposes a simple retinal image segmentation algorithm based on threshold segmentation. Firstly we select a point in the gray image and compares this point and other points for differences in gray value. Through this method, we can determine whether this point is on the blood vessels image. Secondly, we set the threshold to eliminate noise, according to the connectivity of blood vessels. Finally, we obtain the retinal image segmentation results, which demonstrate that this algorithm can effectively extractthe retinal blood vessels distribution, which has a good reference value.

To solve the speed bottleneck of deformable part models in the detection process, this paper proposes a cascade deformable part model with rapid computation of feature pyramids. As the speed of the detection is largely determined by featurecomputation and object location, we propose a two-stage acceleration method. Sparsely-sampled feature pyramids on the scale are firstly utilized to approximate finely-sampled feature pyramids and then combined with the cascade algorithm in the location process. The resultant sequence model is utilized to evaluate parts sequentially so as to rapidly remove most object hypotheses of small possibilities. The experimental results on Pascal VOC 2007 dataset and INRIA dataset demonstrate that the proposed method accelerates the speed of detection with minor loss in detection precision.

An effective and feasible analysis on airport ground operation efficiency and capacity is of great significance for the overall planning of an airport construction project. Utilizing the Beijing New Airport as the test location and applying the simulation software Simmod, the ground operation efficiency and the flight delay time in southward and northward operation directions of the airport is studied in this paper. Based on simulations and taking into account the configuration of terminal building, three apron taxiway operation patterns were proposed, analyzed, and compared through simulation. The simulation results indicate that under the current construction scheme, the Beijing New Airport is expected to run with an average ground delay time of 2.72 minutes for departure flights and 58 seconds for arrival flights in target year, which is of high operation efficiency. Over half of ground delays of departure flights is attributed to takeoff waiting time, another 1/5 happens in the angle area between main buildings. When the three apron taxiways within the angle area operate in two out and one in pattern, ground delay time is minimized.

To accommodate the needs of local mobile communications, the interface of data marts are designed using the data warehouse and ETL2. The interface of the data warehouse is composed by the user interface, the interface traffic, the billing interface, etc. ETL2 is designed by processes involving extracting, FTP, clearing and loading. The results demonstrate that the designs can reduce the impact and stress of the data warehouse, and are important supplements to the extension of the data warehouse.

In order to estimate the spatial orientation of weak signal accurately, DOA estimation algorithm for coherent weak target signal in strong interference environment was proposed. Firstly, this algorithm carried out the decorrelation of the coherent weak signal via modified Toeplitz matrix. Then, the strong interference signal corresponding to the feature vector was removed from the subspace. Finally, DOA estimation of the weak signal was accomplished by making use of MUSIC algorithm. Simulation results indicate that the algorithm can eliminate the influence of strong interference signal effectively.

To ensure the safe navigation of ships, this paper proposes a three-dimensional model for dynamic ship domain by integrating ship wave and turbulent region theories, taking the impact of water and wind into consideration by utilizing quantitative analysis to update the equations of lateral boundary and longitudinal boundary in accordance. Furthermore, data about the experiences of ship pilots is applied to adjust the model with a coefficient correction method. The experimental results demonstrate that the established model has higher scalability compared with existing models. In addition, the model's performance in the experiments also demonstrates that our model is a better model to simulate actual shipping conditions.

Natural human-computer interaction is an active research topic in the field of computer vision with widespread applications and one important application is in computer game. In this paper, we propose a novel method using forearm gesture as the input of the Tetris game via an ordinary camera. We use the particle filter to track the movement of the forearm and thus obtain directional information. Thereafter, the directional information is mapped into the input of the Tetris game. In comparison with existing human-computer interaction (HCI) systems which require expensive equipment such as stereo cameras, our system only requires an ordinary camera. Experimental results demonstrate that our method effectively tracks the movement of a human forearm, and implement real-time control for the Tetris game.

Stochastic dynamic programming (SDP) is widely adopted in a long-term optimal operation of large-scale hydropower systems. In this paper we propose a distributed parallel stochastic dynamic programming algorithm based on Message Passing Interface (MPI) and a peer-to-peer parallel paradigm (DPSDPoM) To deal with the disadvantages of redundancy in communications and memory-consumption during calculation in the peer-to-peer parallel paradigm, we propose a DPSDPoM with multithread algorithm (DPSDPoM-MT) which reduces costs between processes in each machine. The two algorithms are compared through theoptimization scheduling of three reservoirs on Lancang Jiang Dam Cascade using time-elapse and memoryconsumption. Experimental results demonstrate that the improved algorithm can reduce computing time and alleviate memory consumption effectively.

DPOP is an efficient algorithm based on the Depth First Search (DFS) Pseudo-tree for distributed constraint optimization problems in multiagent systems (MAS). DFS Pseudotree is able to help achieve parallelism due to the relative independence of nodes lying in different branches. However, we often get a chain-like pseudo-tree with few branches in our experiments, which greatly impairs the algorithm performance. Therefore we propose a new DPOP algorithm called BFSDPOP which uses Breadth First Search (BFS) Pseudotree as the communication structure. The two advantages are that BFS Pseudo-tree can help the algorithm achieve more parallelism as it has more branches; and BFS Pseudotree shortens the communication path and requires less communication time because the height of a BFS Pseudo-tree is often much lower than that of a DFS Pseudo-tree from the same constraint graph. To overcome the cross edge constraints in BFS Pseudo-tree which can easily result in large utility message size, a method of Cluster Removing is proposed. In the experiment, we compare BFSDPOP with the original DPOP and the result shows that BFSDPOP outperforms the original DPOP in most cases, which demonstrate the excellent attributes that BFS Pseudo-tree has over DFS Pseudo-tree.

In this paper, an improved interframe coding scheme consisting of both the adaptive directional motion search and improved mode selection is proposed to reduce the computational complexity of video coding. The motion direction is divided into four regions, and different multi-hexagon templates are designed for each region to enhance directional search, which can improve the search accuracy. Subsequently, the improved mode decision that integrates the detection of all zero blocks is proposed for further computation reduction. Experimental results demonstrate that the proposed scheme can reduce encoding time by an average of 60% while maintaining almost the same coding efficiency as existing fast interframe coding schemes.

In this paper, fast mode decision strategies based on motion coherence for P-frame coding are proposed to accelerate the video encoding procedure. Firstly, mode prejudgment strategy based on neighborhood macroblocks (MBs) is introduced. According to the coherence of adjacent movement, the left and up MBs in current frame and the current-position and right MBs in previous frame are used to predict the mode of current MB. Secondly, mode combination strategy based on motion coherence is implemented to combine blocks in sub-type with similar motion vectors (MVs) into a block in macro-type. Lastly, the overall fast mode decision scheme is proposed and the performance of the proposed scheme is verified through comparative experimental results using JM reference software.

Taxis equipped with GPS can record their trajectory and generate huge amounts of data. We can analyzethe behavior of taxi drivers and search for similarities and common characteristics in their working patterns. In this paper, we utilize taxi GPS data collected form Tianjin city to analyze taxi drivers’ working pattern. Firstly, we determined taxis’ parking place by detecting stopping points and measured the operating range. Secondly, we studied the taxis’ working pattern by comparing the relationship between three representative locations in taxis’ behavior: parking place, working center and city center. Thirdly, we analyzed the spatial pattern from two perspectives of direction and distance. Lastly, we studied the income efficiency of different taxis from different parking places and working centers. The research results demonstrated that the taxis’ individual mobility behavior has clear similarities: The taxi drivers’ operating pattern has a characteristic of moving toward the city center andmost of the working centers distribute in positions between parking places and the city center. The discoveries are significant for urban public infrastructure construction, government’s management of the taxis and taxi drivers’ strategy selections.

This paper designed an intelligent blind cane system based on multi-dimension environment perception to provide improved protection and safety for the blind. This system utilizes the Arduino platform as the core to connect the hardware, and uses the edge detection algorithm, calibration algorithm (temperature calibration and mean filter), Kalman filteringalgorithm to achieve obstacle detection, GPS positioning, traffic light recognition and other functions..

Utilizing the advantage of increased revisit performance of the low inclination regression orbit; this paper developed the optimization design model considering the constellation configuration, the orbit and the performance, and designedthe constellation configuration optimization design process. Based on Multi-Island Genetic Algorithm (MIGA), the multi-objective optimization design of a satellit constellation with rapid revisit in equal time interval is completed. Performance analysis results for satellite constellation optimized indicated that the mission requirements and design constraints are satisfied, and the rapid revisit to the target in equal time interval is achieved.

When topic evolution is analyzed by the traditional topic model on the micro blog data, the problem of sparse features arises. In this paper, we propose a new topic evolution method of combined BTM model with community discovery technology, which called BTM-LPA method. Firstly, this method uses the BTM model to select feature words and model in different time windows, and construct the co-word network by using the cooccurrence relationship of feature words in each time window; Secondly, through the community discovery technology, we use the LPA algorithm to find communities in coword network, where a community represents a subtopic and the similarity measure of subtopics is transformed into the similarity measure of communities; Lastly, the five evolution types of subtopics, which include birth, extinct, merge, split and development, are defined for correlation analysis, the correlations of the subtopics are then set up according to the semantic similarity and temporal relations and the topic evolution is described through content change of subtopics. The experiment results demonstrate that the proposed method in this paper can effectively reflect the evolution process of the micro blog topic.

Social networks play an important and indispensable role in the internet, especially for governments and organizations. Social networks are information diffusion networks with users as the node and the relationships between users functioning as the vehicle. In this paper we propose three different type of algorithms to compute the user influence based on users' behavior of forwarding micro blogs and the symbol of @ in the micro blogs. We evaluate the effectiveness of the algorithm by comparing the results of our algorithms and the trained data in the dataset and the results demonstrate promising performance.

The identification of essential genes is important for understanding cellular mechanisms and determining potential drug targets. Neural Network methods have been widely used in biological fields and in this paper, we construct a Back Propagation Neural Network (BPNN) model to predict essential gen. Three evaluation methods were designed to assess the predicting effect: (1) Self-test on a single organism (Escherichia coli). (2) A BPNN model trained by a single organism (E. coli) is used to predict the essential genes of other 30 organisms. (3) A BPNN model trained by 30 organisms is used to predi the essential genes of another organism (E. coli). In this model, 57 sequence-based features were used to screen the key feature(s) related to gene essentiality and we proposed a method based on Principal Component Analysis, which reduced the quantity of features from 57 to 35 and the prediction performance remained stable. These screened features could be used as key features in computational analysis and biological experiments.

To overcome speaker segmentation in the walkie-talkie radio system’s speech, we propose a novel method of recognizing the acoustic presence of push-to-talk clicks. Firstly, a typical speaker segmentation system is realized by GMM with BIC criteria. By moving the time window with different scales, speaker-change points can be determined and the best performance value of F is 65.47%. Secondly, a detector is used in acoustic analysis to locate the acoustic event of push-to-talk click. Lastly, a fusion scheme is designed to take advantage of the both results. The F value of the proposed method can be boosted up to 77.18% with a relative increment of 17.88%. In addition, the recall rate and accuracy rate are improved by 20.01% and 16.07% respectively.

To determine the similarity measure of news articles, this paper proposes a hybrid approach based on the dictionary and research done by previous studies. We identify representative nouns from news articles, and combine the word similarity metric into the text metric, which helps the news provider to filter out similar articles to the subscriber. The experiments are conducted to examine a set of press releases and news articles, while the results demonstrate that our proposed method outperforms the TF-IDF-based approach.

In order to effectively replace the background of micro moving targets in the video sequences, it is important to solve the problem of micro moving targets extraction. We propose a new algorithm of micro moving target detection in video images. Firstly, Adaboost was used to train skin samples to obtain the skin color data and determine the threshold range of skin color. Secondly, in order to obtain the rough binary images, we fused the binary images of skin detection with the binary images of micro targets by the Otsu. Thirdly, morphological processing and the holes filling processing were used to overcome the rough binary images, in order to obtain the complete binary templates. Lastly, the algorithm successfully accomplished the background replacement of the video sequences. The experiments demonstrate that the proposed algorithm in this paper is effective, which can segment micro moving targets accurately and complete background replacement.

The establishment of a Codebook background model in RGB color space results in high computational complexity, low segmentation quality and poor immunity power. We proposed the improved brightness range Codebook algorithm under YUV space to overcome the problem of irrational brightness range definition of the algorithm, as well as the background updating method which combined a two-layer Codebook model with a short-sliding window. Firstly, the Codebook background model was built under YUV color space based on the characteristics of luminance and chrominance separation. Secondly, the brightness range was redefined using the average of the Y component of the codeword to tackle the problem of the irrational brightness range definition of the background model. Lastly, a two-layer Codebook model and a short-sliding window were combined to update the background in the process of moving object detection in order to improve the antijamming capability. The experiments demonstrated that the improved brightness range can be updated gradually with background changes. Comparing with the original algorithm, this method can update the background effectively, and improve the detection accuracy.

Big Data technology has experienced rapid development in recent years and has demonstrated its effectiveness when implemented inclinical diagnostic decision support. Current research of big data focuses largely on data storage architecture, the exploration on big clinical data mining and decision making strategy is relatively less. This paper discusses the parallel big medical data mining method under Spark architecture and proposes the optimized support vector machine (PCSVM) algorithm to provide qualitative data for clinical diagnostic decision support.

The scanner of three-dimensional reconstruction has become popular in the common practice of quality measuring, quality control, and cultural heritage et al. In this paper, we presented an algorithm to produce high-quality of photorealistic textured model, which aligned and integrated range images into a common system, reconstructed the geometric shape of object; In the next step, the captured color-images are back-projected onto the reconstructed surface and removed the artifact with a composite weighted strategy. The experimental result shows our presented method can produce high quality photo-realistic results with a low computational cost.

We present a deep neural decision tree based on Spark - a novel approach that unifies classification tree with the representation learning functionality known from deep belief network and runs on the Spark to dramatically decrease the training time of the model. Firstly, we introduce a stochastic, differentiable, and therefore back-propagation compatible version of decision tree, guiding the representation learning in lower layers of deep belief networks. Furthermore, we make an parallel implementation for the decision tree back-propagation method based on Spark. We show experimental results on the Sloan Digital Sky Survey Data Release 7(SDSS-DR7) star/galaxy set and find out it scales well with the cluster size and have on-par or superior results in the classification accuracy when compared to state-of-the-art model which are used on the data set.

In order to investigate the actuality and demand of transmission line tower foundation, this paper proposes an engineering design management software which comprises “input of foundation parameters”, “design method selection”, “design parameters selection”, “foundation type selection”, “output of design calculation sheets”, “output of working drawing” using Visual Basic and Visual C++ developing environment and AutoCAD ObjectARX and VBA Office programming technology. The software can precede foundation parameters calls, inquiry of foundation type selection, foundation optimal design and design result batch output. This software provides a reasonable, precise, fast and economic design tools for the designers, thus to realize the overhead transmission lines tower system, basic engineering design innovation, safety and economic goals.

The study of code reverse engineering requires the integration of disassembly and decompiling key technologies. As independent adoption of disassembler and decompiler causes decompiling input restriction, e-CRT is implemented in this paper to overcome this issue. Its sound effects are presented not only through theoretical foundation but also in a specific test case of the smart meter embedded device

The traveling salesman problem (TSP) involves the question: Given a list of cities and the distances between each pair of cities, what is the shortest possible route that visits each city exactly once and returns to the origin city? Additionally, it belongs to the class of NP-complete problems. It has been proved that bionic intelligence algorithms are effective and efficient, with respect to the traditional methods for solving TSP. Artificial Bee Colony (ABC) algorithm is a swarm intelligence optimization algorithm based on the foraging behavior of honey bee swarm. In this paper, we employ the artificial bee colony algorithm to solve TSP, present specific solutions of artificial bee colony algorithm, and conduct a simulation experiment to solve TSP. The results demonstrate that the algorithm can efficiently and rapidly search for optimal or near-optimal solutions.

The improved resampling algorithms commonly in particle filter (PF), increase particles’ diversity by making new particles with various methods, and thus improve PF’s accuracy. However, they also increase the distance of particle probability distribution before resampling and reduce theactual estimation accuracy. To solve this problem, thispaper proposes an improved Gaussian resampling(IGR) algorithm, based on Gaussian Resampling (GR) Algorithm. Under the premise of maintaining the diversity of particles, we enable new particles to contain part of the low-weighted particles’ information by conducting proper linear combination with low-weighted particles. Simulation experiments conducted on single variable non-growth model suggest that the improved algorithm reduces the particles’ Kullback-Leibler(K-L) distance and improves the final tracking accuracy of PF.

In this paper, the fully discrete method is applied for the eigensolutions of the Laplace equationon smooth closed boundaries. The fully discrete method mainly consists of two levels of numerical quadrature: the trapezoidal rule for theintegrals including the weakly singularity, and the discrete inner product for the outer integrals. The convergence and error results of the fully discrete method for Steklov eigenvalue problems are provided. Thereafter, the numerical examples demonstrate the efficiency of the fully discrete method.

In this paper, the set of some points, derived at which the degree of a Boolean function can be reduced by at least i, is defined as . We show the set of i-DPs (i-differential point) is closed to the operation XOR, and the origin i-DP still exists after taking a derivative on the Boolean function at a non-2-DP. In addition, ann-variable non-linear Boolean function of degree d has at most 2^n-d– 1 2-DPs. We prove that an n-bit cipher whose n component functions have common terms of degree n – 1, is not IND-CPA secure. We also show that an n-bit cipher with low degree d can be distinguished with obvious advantage and data complexity 2^d+1.

In this paper, we select 4 indexes that are commonly used to measure the degree of the non-singular of the matrix: the minimum norm of eigenvalue, the minimum norm of singular value, the condition number and the absolute value of the determinant. The Pearson correlation coefficient was calculated for every 2 pairs for the 4 indicators of the random non-singular matrix so as to compare the changes between them and to explore the correlation between these indicators. The experimental results demonstrate that there is a good linear correlation between each pair of the 4 indexes.

Service-oriented computing is emerging method to develop extensible computing systems which evolved from component-based software engineering. The complex web services can be composited by simpler, loosely coupled, reusable web services. Reliability is an important factor for choosing, ranking and compositing Web services. However, existing analysis methods of composite web services are rare and majority of them are proposed for component-based software reliability analysis. In this paper, a reliability analysis method for composite Web services is proposed. The architecture of the composite Web service is characterized as a tree which is transformed from its business process execution language (BPEL) description. In addition, several examples are illustrated to demonstrate the effectiveness of the proposed analysis method.

As existing similarity algorithms are hampered by low accuracy and recall rates, a comprehensive ontology-based weighted similarity algorithm was proposed for case retrieval. Taking into account the various attributes of the case, different attribute calculation methods were used and each attribute was weighted in accordance and the similarity of the case was calculated. The similarity algorithm was validated by three psychological counseling cases. By comparing the traditional similarity algorithm based on the nearest distance and the weighted similarity algorithm, this paper demonstrates that the weighted similarity algorithm exhibit higher accuracy.

In the RBAC model, a resource is represented by an object. As there is close connectivity between objects and operation, the resources accessing control authorization is inconvenient. This paper extends the RBAC object model to the resource model while the resource elements, abstract resources and resource bundles are defined in the resource model. The RBAC operating model is extended and the concept of abstract operations and the operating bundles are proposed. In addition, the relationship between resources and operations are discussed and the grading of the resource and role is proposed.

This paper discusses the development of soft-parts to achieve software reuse and improve the efficiency of software development. Based on the JAVA platform, we researched the common problems of soft-parts and demonstrated the fundamental steps in constructing applied systems through this system. The tests demonstrate that the application of softparts can significantly improve the efficiency of software development.

An improved 2–D coupled map lattice model based on radial basis function is proposed fit the evolution of an actual spatio-temporal system. A real signal was adopted and by using different coupled map lattice (CML) models, different predicted signals were obtained. The correlation coefficients between the real signal and the predicted signal indicated the fitting capabilities of the models to real system. Simulation results demonstrated that the proposed method can effectively fit the evolution of an actualspatio-temporal system.

In order to solve the problem of a lack of accuracy in absolute single point positioning in certain fields of application, a design scheme of software pseudorange differential positioning system (BDS) is proposed. To tackle the problem of original measurement pseudorange smoothing and denoising, this paper proposes an improved integral Doppler smoothing pseudorange algorithm. An improved weighted least square algorithm is introduced to reduce the influence of measurement noise on the localization to overcome the limitations of the traditional least square algorithm. The experimental results demonstrate that the proposed algorithm can provide a stable planar position accuracy of 2∼5m which is three times higher than single point positioning.

Traditional beamforming algorithms have relatively poor performance in conditions of small snapshot numbers, high signal-to-noise-ratio and coherent sources. To address this issue, an oblique projection-based beamforming algorithm is modified in this paper. In this algorithm, the oblique projector is utilized to eliminate the noise of the array input data and enhance the robustness of algorithm. Furthermore, the transformation-based linear constraint matrix may eliminate the interferences, and Chebyshev window function is utilized to suppress the side-lobe level. Simulation results demonstrate that the modified algorithm has good robustness and performance.

This paper proposes a new direction finding system consisting of a transmitter and two receivers. The phase differences of different frequency signals received by the two receivers are used to resolve interferometer phase ambiguity in this method and the direction is obtained through the phase differences obtained. The simulation results demonstrate the effectiveness of this method.

EDMA (Enhanced Direct memory access) is an important method to transmit data in Digital Signal Processer (DSP) for high speed and real-time processing of digital signal, such as digital image processing, radar signal recognition. However existing DMAs do not use data pre-computation and they are incapable of satisfying the increasing demands of signal processing, target recognition or scientific computing applications. In this paper, we propose a new enhanced DMA design with data pre-computing for digital image processing, which provides more efficient data transfers in modern DSP by integrating several useful arithmetic units for data preprocessing. In comparison to traditional EDMA, the proposed design reduces the DSP’s burden of simple calculations and increases the efficiency of data transmission for improved performance..

Similarity measurement plays an important role in the classification of short text. However, traditional text similarity measures fail to achieve a high accuracy because the sparse features in short text. In this paper, we propose a new method based on the different number of hidden topics, which are derived through well-known topic models such as Latent Dirichlet Allocation (LDA). We obtain the related topics, and integrate the topics with the features of short text in order to decrease the sparseness and improve the word co-occurrences. Numerous experiments were conducted on the open data set (Wikipedia dataset) and the results demonstrated that our proposed method improves classification accuracy by 14.03% on the k-nearest neighbors algorithm (KNN). This indicates that our method outperforms other state-of-the-art methods which do not utilize hidden topics and validates that the method is effective.

Identifying the effective connectivity of human brain is much crucial for understanding the cognition and brain diseases. Transfer Entropy (TE) offers a model-free approach to detect directed information flow between nodes in brain network and is inherently nonlinear. In this paper we apply TE to the detection of the effective connectivity of brain network with 998 ROIs (Regions of interest, or nodes). With the resting state functional MRI (rs-fMRI) BOLD time series we obtain the strength of directed effective connection and analyze its relationship with the structural connections.

This paper proposes the design process of a DDS signal processing unit based on the SOPC technology. The main hardware is implemented by the NiosII soft core and the corresponding interface module on FPGA chip. The chip EP2C35 is used to control and deal with the signal which comes from the DDS module. Implementation of custom instruction hardware in SOPC Builder introduces the RS coding function. Comparison of software operation results demonstrated that our designed system has faster development and a high performance to price ratio, thus it can be widely applied in many fields of engineering and technology.

This paper proposes a DSP fault tolerant approach based on loop optimization known as DSP Loop Optimization Approach (DLOA). DLOA reduces the performance overhead incurred by traditional fault tolerance techniques while maintaining their fault tolerance capabilities. DLOA delays the fault tolerance latency between errors detecting and errors handling to scheduling the software pipeline, increasing performance significantly. The performance experiments and the ion irradiation experiments in the Heavy Ion Research Facility in Lanzhou (HIRFL) demonstrated that DLOA used in SWIFT achieved a 6.2 times average speedup and with its fault tolerance ability unaffected.

An assisted Teaching Information Publishing System (TIPS) based on glossy screen is designed and developed in this paper for displaying teaching information dynamically. This system is composed of one information publishing server and more than one information display client terminals. With C/S program mode and computer network communication, the teaching information can be published to each display terminal accurately and timely. By integrating human perception sensors installed on the display terminals, the observer’s position in front of the glossy mirror is confirmed, then the size of the display contents will be adjusted according to the distance between the mirror and the observer. When there is no human or the relative distance is distant, that’s to say, it is beyond the set range, the glossy screen will show nothing and shut off its display function for achieving to save power and improve its life. At this time, this glossy screen as an ordinary mirror will be used for checking the teachers and students’ dress whether they are proper or not. As a new way of display, it gives the teachers, students as well as the other audiences new experience, thus makes a large improvement for them to receive published information actively.

In order to overcome the shortcoming of small standoff distance of the submerged cavitating water jet, this paper proposes a new artificial submerged cavitating water jet nozzle. Fluent is used to perform the inner flow field simulation of the nozzle and several nozzles’ structural parameters are optimized through the total volume integral of vapor. The results demonstrate that the flow accelerates in the convergent section of the inner nozzle and expands in the diffuser. Additionally, cavitation takes place initially in the diffuser and small cylindrical section of the outer nozzle and the nozzles’ structural parameters have significant effects on cavitation. The cavitation quality is optimal when the outlet diameter of outer nozzle is 10 mm and the distance from the inner nozzle’s outlet to the outer nozzle’s convergent is 6 mm.

This paper proposes a cloud computing based fault diagnosis scheme for radar systems. Based on the gradation diagnosis model and fault separation tree, diagnosis knowledge is represented by an integrated framework; and the direct inference, reverse inference and hybrid inference diagnosis strategies are carried out. Site test results demonstrated that the diagnosis system improves the fault diagnosis accuracy and speed of modern radar fault diagnosis.

In this paper, the actual conditions of the hydraulic torque converter production workshop are analyzed, and the overall architecture, key technology and function module of MES in the production workshop of hydraulic torque converter are present. The use of MES greatly improves the efficiency of the hydraulic torque converter production.

To solve the time-consuming problem of K-SVD, a dictionary learning algorithm based on clustering theory is proposed. The proposed algorithm learns incremental dictionary from an incremental training sample set, then leverages clustering theory to combine the initial dictionary and incremental dictionary. Thus it provides a two way optimization for the entire training sample set and significantly increase the efficiency of the K-SVD dictionary learning algorithm for big sample set learning. The image super resolution reconstruction experiment demonstrates that the proposed algorithm exhibits the capability for incremental learning.

This paper discusses the basic principle of the neuron PID controller for the brushless DC motor. Unlike the classical PID controller, the neuron utilizes the deviation, integration and derivative values as inputs and utilizes the proportional, integral and differential gain as the weight of the neuron, thus making the neuron PID controller adaptive. Experimental results demonstrate that the neuron PID controller is a better control method as compared to the classic PID controller.

The Electroencephologram (EEG) experiments were designed in this paper to emulate our daily visual working memory task and demonstrate the effect of image target numbers and background textures on our brains’ visual working memory. This paper discovered that there is a possibly of implementing a neural network to predict memoryrelated brain activity in a visual working memory experiment, where participants were presented with images of different target item numbers and asked to remember as many target objects as possible. Both the Multi-Layer Perceptron (MLP) network and support vector machine (SVM) were used as training methods for the prediction. The prediction results are consistent with the actual EEG power variation observed in the experiment, which demonstrate the effect of target item number and background texture on the level of difficulty in image memorization.

Traditional FPGA placement algorithms based on simulated annealing is time-consuming and thus we have proposed a parallel FPGA timing-driven placement algorithm using OpenMP + STM programming method. In this paper, we distribute swaps to multithreads by OpenMP and protect the shared memory using software transactional memory. An improved timing optimization algorithm is also added in the transaction. Experimental results on MCNC benchmarks demonstrate that our algorithm achieves a speedup of 1.6x and scales well with the increasing of threads. It also reduces the critical path delay by an average of 4.2%.

The cultivation of programming skills is the essential goal of professional education in application-oriented universities. To address the problems in the professional education of computer science undergraduate students in the two courses, C Programming Language and Data Structure, the paper introduces the current knowledge structure of the two courses. It then reconstructs the modularized education content focused on the data object, and proposes the curricular structure of Teach after Study. It can improve the practical teaching system that is currently driven by assignments and projects, optimize the assessment system, and eventually lead to the curriculum integration of C Programming Language & Data Structure and pedagogy based on the cultivation of programming skills.

Currently, there is an increasing dependence on the educational administration system in colleges and universities. As the educational administration system supports the school teaching resources, the massive access pressure during course selection is a problem that is becoming increasingly prominent. We propose the construction of a cloud platform to pool all physical servers and the setting up of virtual machines. During course selection, the cloud platform can increase or decrease the number of virtual machines automatically through the load elastic flexible configuration and the monitoring of parameters such as the number of requests and CPU usage. This greatly improves the efficiency of course selectionand student satisfaction while significantly reducing server hardware expenditure and raising the efficiency of maintenance and operations.

This paper proposes an integral predictive control strategy, a hierarchical scheme consisting of the Runge-kutta controller and PID controller to stabilize the rotational movements of the AR Drone. The Runge-kutta algorithm can predict states in advance and hence we utilize this robust algorithm to build the Runge-kutta controller. The control strategy combining both the Runge-kutta controller and PID controller can be considered as a Robust PID controller, which improves the drone’s robust performance especially when flying at high speed. The controller reduces the positioning errorin conditions of continuous disturbances in the operation environment. The experimental results demonstrate that the Robust PID controller possess good accuracy and robustness in drone control for both indoor and outdoor environments.

Current methods for identifying essential proteins based on protein-protein interaction network do not provide sufficient information about the biological functions of proteins. In order to solve this problem, we introduce the use of protein complex information and propose an algorithm named CRW. CRW is based on the random walks model and integrates edge clustering coefficients in protein-protein interaction network in order to predict essential proteins. The experiment results demonstrate that the number of essential proteins identified by CRW exceeds the number identified by the five centrality measure methods.

In this paper, we use the PID control algorithm based on the calculative optimization method to control the concentration of the 9-dehydro-17-hydro-andrographolide administered to the liver of SD rats. Through data processing and with the assistance of the polynomial fitting function in MATLAB, we obtained a fitting analysis by using the 9-dehydro-17-hydro-andrographolide distribution data in SD rats’ liver tissues and achieving a close fitting effect. Based on the fitting function obtained, we can not only design the PID control system and calculate the PID controller parameters that content of the sensitivity index, but also utilize the PID algorithm to attain the effect of the expected model. The mathematical model that we developed is clear and accurate, it can reflect the dynamic changes law of the drug in the body, from which, we can observe and analyzes the input-output curves to draw accurate conclusions. Simulation results demonstrate that the control algorithm has relatively high adaptability and robustness, and we can obtain better control effects.

This paper discusses the evaluation method of mobile internet social apps through a humanist perspective. We analyzed classic humanlogy theories such as Marxist practical human thought and demand theory, Freud’s sub consciousness theory, Sartre’s freedom to pursue ideas, Fromm’s interpersonal and social demand, Marcuse creative lust theory and Maslow’s hierarchy of needs in the growth of the existing research achievements, etc. By integrating the interaction design theory with related digital art design theory, we carried out a series of experiments to explore this new evaluation method for social applications. Research conducted in this paper based on the study of Social applications DPCF (Demand Production Consume Feeling) model and the SFT (Social Feeling Tools) Social application evaluation model, provides new methods to optimize the design of social applications.

In this paper, we propose an approximate method for global illumination. We build an occluded estimator based on an algorithm of parallel line sweep for shadow maps. The virtual point lights are classified and a method is proposed to generate irradiance estimator (IE) by azimuthally analysis of normal. For each pixel rendering, we sample the IE around the pixel to fast approximate the percentage of irradiance to the point in the scene. The experiments demonstrate that our algorithm is an effective method of approximate global illumination.

The following sections is included:

• Author Index