Abstract:For the problem that failure nodes lead to inaccurate topology inference results in the power line communication, an advanced Tomography-based topology inference algorithm has been proposeed. The algorithm introduces node instability degree and intra-cluster hierarchy to handle with the failure nodes. Firstly the node instability degree is determined according to the failure node communication. And then the intra-cluster hierarchy is used to make the failure node only be a member of the cluster instead of a cluster head. Thus it improves the accuracy of the topology inference result. Experiments show that the algorithm achieves positioning and processing the failure node in the topology inference stage, and effectively improves the topology inference accuracy. As fault prediction has the characteristics of uncertainty, we design a method for fault prediction, which combines fuzzy mathematics membership function with particle filter algorithm to predict fault. The new method uses particle filter algorithm to calculate the future state of the device operation, and then designs the normal membership function and the abnormal membership function of the device operation state, calculates and compares the value of the normal and abnormal membership function by using the calculated results and based on the comparison result to predict potential failure. The feasibility of the proposed method is verified by experiments, which can predict the failure of the system in time. The traditional authentication technology showed some deficiencies in cloud computing, in order to decrease intrusion of untrustworthy user, this paper discusses a scheme for user behavior authentication in cloud computing environment which combines traditional authentication and behavior authentication. The paper presents the creation of the user behavior authentication set. The paper presents the realization process of the whole mechanism. The paper proposes Bayesian network model to predict the level of user behavior authentication based on the combination of the evidence of past transaction and real-time user behavior information. We give an example to demonstrate the effectiveness of the prediction model. This research is of theoretically and practically significant for enhancing user authentication and decreasing intrusion of untrustworthy user effectively. The accuracy of the hand feature points location directly affects the correctness of the hand matching. On the basis of the existing hand features location, this paper proposes a feature points location algorithm, which is based on partial blocks scanning. The algorithm starts from the middle finger peak which is the most easy to locate. Then we determine the coordinates of the rest fingers peak points and valley points, step by step according to a fixed order recursive formula block based on the coordinate of the middle finger peak. Then, we extract the hand feature vectors based on the determined feature point and use matching algorithm based on the hand feature vectors to match. Experimental results show that the success rate of the algorithm's feature points is up to 93% and the accuracy rate of the hand shape recognition is up to 80.5%. This indicates that the algorithm can accurately locate the hand feature points and the algorithm is feasible. The 3D model animation plays an important role in the digital design and application. Though many scholars have got research achievements of 3D model animation, how to reproduce the national dance by 3D digital technology is a challenging problem now. This paper uses the motion capture technology to display 3D digital national dance and compared it with the method only based on the 3D model. Firstly, the motion capture equipment gets human motion data file and then creates 3D model which included skeleton, skin and weight work in Maya. By MotionBuilder, 3D model was combined with motion capture data. In this paper the real dance of 13 different nationalities finally is shown by virtual characters at virtual scene. RP(Relying Party) downloads signed objects in RPKI(Resource Public Key Infrastructure) and processes those objects into authorised relations between IP addresses and AS(Autonomous System), which is used to guide the BGP routing. The current RP uses rsync to realize the synchronization, but rsync(Remote Sync) synchronization algorithm does not take the characteristics of the files (directorise) in the RPKI into account. So the synchronization is not efficient. Through the analysis and combining with the characteristics of the files (directories) in the RPKI, this paper designs and realizes a RPKI repository synchronization tool named htsync which is based on ordered hash tree. The experimental results show that, compared with rsync, htsync transmits less data and costs less time during synchronization. In three designed experimental scenario, average speedup ratios of synchronization time are 38.70%, 30.13% and 3.63%, effectively reduce the consumption of time and resources. First of all,this paper analyzes the basic steps of dark channel prior algorithm, and then proposes a method to improve image quality under special scene image restoration algorithm. The algorithm works on special scene image through reversion operation firstly, and then lets improved dark channel prior algorithm use for inverting image to make the quality of the special image be improved on using the conclusion that an reversed special scene image has the similar features with a haze image. In this paper, we will optimize the parameters of improved dark channel prior algorithm and deal with the sky in the image area and the non-sky area in separate, which is very good visual effect and also reduces the complexity of the algorithm. Collaborative Filtering(CF) is one of the most successful approaches for building recommender system,it uses the known preferences of a group of users to make predictions of unknown preferences of other users. The matrix factorization models which can profile both users and items latent factors directly,and the neighborhood models which can analyze similarities between users and items are current research focuses.A method of merging both matrix factorization models and neighborhood models is proposed, which can make further accuracy improvements. The experiment results show that this method is correct and feasible. In this paper, we proposed a vertical handover algorithm based on speed prediction for high-mobility users, such as the scenario of the highway. The algorithm predicts the RSS(received signal strength) of WIMAX and LTE networks in the next time based on the velocity obtained by APA(acceleration prediction algorithm). Velocities are predicted by APA. Then velocities multiply their timeslots to get moving distance respectively. Then current positions are added with moving distances, thus gets next time's positions. Finally distance between AP (access point)and MN(moving node) are calculated by the relationship between distance and present RSS. In the end, the next time's RSS are predicted. Then an effective handover algorithm is proposed using the predicted RSS aiming to reduce the handover times and shorten handover latency. The simulations results indicate that, the proposed algorithm can reduce about 10% handover times and reduce handover latency with sufficient RSS is guaranteed in communication, comparing to the traditional dwell time algorithm. The compression of the GPS trajectories in a conventional way is greatly different with the original one for ignoring the speed information, the direction of GPS points and the shape of tracks. In this paper, we focus on saving the speed, the direction information and shape features of tracks on the premise of keeping the compression ratio. The Algorithm in this article is based on the road network information, OW(Opening Window) algorithm, the core point algorithm and the stay point algorithm, then presents a online compression algorithm which is capable of saving the temporal characteristics of GPS tracks. Experimental results show that the compression algorithm compareing to the existing compression algorithms, ensurs the compression ratio while improving to retain the temporal characteristics of the track. In this paper, a traffic flow criterion is established which is based on extracting the features of the traffic flow and dividing its state. Combined with spatial and temporal dimensions, a new method is proposed to judge the congestion level of the current road according to the regional density and the time occupancy as the features of the traffic flow state. The simulation shows that this method has better accuracy and objectivity. In order to solve the problem of inaccurate priority judgment and complex operation interface and other issues of the existing image restoration algorithm based on sample block, the thesis introduces the concept of the body structure and takes advantage of the graphical user interface (GUI) to operate. The algorithm gets the body structure through artificial selection of repaired area, the Gaussian filter, wavelet transformation and edge detection on the platform of Matlab7.11. Then determine the priority, and finally use the sample block repair algorithm to update self-confidence to finish the completion of restoration. According to the experimental results, the PSNR increases by 10.5~11.07% after restoration,in addition, the algorithm operates simply and will get wide range of applications. The features of ZigBee make it suitable for the application of data acquisition system. A method which is designed for the ZigBee-based data acquisition system is proposed. Most traditional positioning based on ZigBee uses wireless signal loss model to locate. The work is affected by the impact of the surrounding environment. If we take experience value in reality and sets it as parameters in the model, both the positioning accuracy and the universal of this method are not in ideal. With Curve Fitting method fitting the relationship between RSSI and distance, then RSSI triangle centroid algorithm is used to calculate the coordinates of the nodes to be measured. Actual experimental result shows that the algorithm can improve the positioning accuracy. Network coding technology has obvious advantages for improving networks throughput,balancing network load,improving bandwidth availability ratio, enhancing robustness of networks, but it can't resist against pollution attacks directly. Recently, scholars propose the signature scheme based on homomorphic Hash function, which could better detect pollution attacks, but it is difficult to locate the contaminated node. This paper proposes a network coding scheme based on the digital signature by combining with the advantages of both. It can not only resist pollution attack, but also can effectively identify position of attack source. Thus, it reduces the impact of pollution attacks to the network and enhances the robustness of the network. Aiming at the insufficience of traditional wavelet soft and hard threshold functions and some of the threshold functions in existing literatures, and to propose an improved wavelet threshold denoising algorithm. A new threshold function is presented, which has better smoothness and changes with the changes of wavelet decomposition scale, so it has good adaptability. Besides, it improves denoising stability as a result of that there is no uncertain parameters in the function. Compared with traditional soft and hard threshold functions, that uses of the improved threshold function denoising, signals' SNR and MSE are better. Wavelet threshold denoising based on improved wavelet threshold shows an excellent effect, which has good promotional value. Tracking-Learning-Detection (TLD) is a kind of long-term visual tracking algorithm which receiveds wide attention in recent years. In order to improve the running speed of this algorithm, a novel algorithm named Accelerated TLD (ATLD) is proposed in this paper. Two aspects of improvements were made in original TLD algorithm. The improvement includes as follows: using a grey prediction model in the detection module for estimating the location of the target and setting a detection area; applying an image indexing method based on normalized cross correlation (NCC) distance to manage the positive and negative sample set. And on this basis, the multiple targets tracking algorithm is realized. Through experiments, the ATLD algorithm, the original TLD algorithm and other two recent improved TLD algorithm are compared. The experimental results show that the ATLD algorithm runs faster on the premise of ensuring the accuracy. Project cost forecasting is a key point in the research on project management, in view of support vector machine parameter optimization problem in project cost forecasting, a new project cost forecasting model (IPSO-SVM) is proposed, which is based on the improved particle swarm optimizing supporting vector machine. Firstly, project cost data is collected and processed, and then support vector machine is used to learn for training samples in which improved particle swarm algorithm is used to optimize kernel function parameters of support vector machine, At last, the simulation experiment is used to test the performance of project cost forecasting by using Matlab 2012. The experimental results show that IPSO-SVM can effectively improve the forecasting accuracy of project cost, and the forecasting results have some practical application values. Disaster rescue needs rapid transportation of both supplies and staff. The abrupt disasters often affect the traffic states, and research on the real-time fastest path under dynamic road conditions has important economic and social values. Aiming at dealing with the abrupt and frequent road variation after disaster, a fast algorithm of real-time Fastest Traffic Path(ARFTP) is proposed, which reduces the amount of re-calculation on the nodes and the sub-paths by classifying the nodes into different types and filtering them, then putting them into corresponding calculation modules, so as to avoid redundant calculations. When the vehicles are on the original shortest path to the disaster area and receive real-time road variation information, the up-to-date fastest shortest path can be calculated by ARFTP rapidly. Experiments have been made to test the correctness and efficiency of ARFTP with many cases of road cases, and the results proved its effect in improving the efficiency which has some guidance significance to disaster relief transportation. For the issues that the photographic image (PIM) and computer generated images (PRCG) identification scheme have features of poor generality and high dimension in image forensics, an image forensics scheme base on forecast error variance analysis in color filter array (CFA) interpolation is proposed. First, the Fourier spectrum of prediction error variance of CFA interpolation is analyzed, and the PIM and PRCG are distinguished according to whether there is a distinct periodic peak phenomenon. Then, the periodic peak model is analyzed, and the source of PIM is identified according to the peak value features. Finally, experiments have been done on natural images from Columbia University and computer generated image database ADVENT. Experimental results show that the proposed scheme can accurately distinguish between PIM and PRCG, and the recognition rate of the PIM source devices (Canon, Nikon and SONY) reached 93%. Focused on the issue that software version upgrades frequently, test cycle time is compressed constantly, and the test workload gets heavy, with the combination of a enterprise application software,and based on the QTP technology platform builds a test automation framework. Firstly, this paper designs a automated testing framework for enterprise application software by understanding of the working principle of QTP and the characteristics of enterprise application software; Secondly, through the design of test case, writeing test scripts, strengthening the test script and running automated test, realizing the software automated test. Practice shows that automated testing is more suitable for regression testing because of the time-consuming of automated testing is 15% quicker than manual testing, through the useing of automated testing framework, solves the completeing a large number of test cases covering problems in a short time, guarantees the quality of released software, improves the test efficiency. In this paper, a new type of SQL Injection attack through HTTP Headers is studied. Through analysising an example of the SQL Injection attack, the principle of the new type of SQL Injection attack is revealed, and the defense for the new type of SQL Injection attack is proposed. A defense model is established via such means as the IP filtering, data validation and machine learning, and this model has such advantages as low invasive, easy realization, high availability and strong expandability. Distributed database HBase has the greater advantage than traditional relational database in large scale data loading but there is also a lot of optimization space. We build HBase environment based on the Hadoop distributed platform, and optimize self-defining data loading algorithm. Firstly, this paper analysis the HBase underlying data store, experiments work out that data loading methods of HBase are insufficient in efficiency and flexibility. Furthermore, it proposes self-defining parallel data loading algorithm, and optimizes the cluster. The experimental results show that the optimized self-defining parallel data loading method can give full play to the cluster performance, has good loading efficiency and data operational capacity. For the issues that the building façade maintenance robot (BFMR) is easily affected by environment and vibration, and poor operation stability, a kind of security and stability control scheme is proposed. It uses the rail brake system to suppress the shock in the docking process of vertical and horizontal robot, and a re-leveling process is conducted to compensate the gap which is equal to the positioning error between the built-in transom rail of the robot and the transom rail of the building. In addition, the proposed vibration suppression system is used to control of wire rope vibration acceleration according to the state estimation of the wire rope dynamic properties. This system suppresses the vibration produced by environmental noise, improves the vertical movement of the vertical robot stability and reliability. The experimental results show that, the method can effectively reduce the influence of vibration. Genetic algorithm and Particle Swarm Optimization algorithm with strong search capability have a very wide range of applications in the optimization problem. This paper focuses on approximate solutions of ordinary differential equations and LP solutions, based on genetic algorithm and particle swarm algorithms, a comparison and analysis of the efficiency of two kinds of optimization problems is made. We then fix other parameters but adjust the particle population, in the purpose to compare optimization capability of GA and PSO in approximate solutions of differential equation and the LP problem. This paper puts forward a mobile recommendersystem model based on collaborative filtering algorithm under the mobile circumstances. We aim at the insufficient degree in integration of the existing algorithm and the instant context, and design a circumstance oriented collaborative filtering recommendation system framework model based on the traditional collaborative filtering algorithm. We use MAE and F1 evaluation index to evaluate the commendation quality between our proposed model, pre-filteringmodel and the general context modeling. Through a series of experiments, it shows that the proposed scheme has the better performance in recommendation quality. An improved image magnification adaptive algorithm was proposed on the basis of seam carving algorithm which may appears the situation that the first k small pixels share the same points. It finds the newly adding position of pixels by some strategies through judging the sharing situation and its sharing rate, which makes the addition of new pixels in non-important area of visual attention more reasonable. The manual intervention to select the important area of visual attention is implemented to avoid perpetual object becoming warped and deformed by wrong judgment itself. Experiments show that the improved algorithms get better effects. In recent years, web usage mining has become a new hotspot in the field of data mining. From the web logs which record information of a large number of network user's behavior, web usage mining discovers the characteristics and potential user access law. This paper uses many real running dates of college homepage. Aiming at running log files, we carry out a comprehensive analysis by using the web mining. Analyzing the interest measure of user to the information content. By using the user access to the page, the system can calculate the user data level of interest on each page, and thereby improving the content and layout of the site. An efficient and reliable data resource authorization policy is the key of effective management, and there is no systemic study on how to make an evaluation on its efficiency and reliability. Firstly, QCC (Quantitative Cost Calculation Method) Method is proposed. Compared to the current methods, QCC makes a distinction between the wrong authorization judgment and wrong refused authorization judgment. And then a policy evaluation mechanism based on QCC is put forward. At last, by carrying out real policy evaluation analysis, we verify the effectiveness of the mechanism in improving the work efficiency with the real log data of an integrated network management system. Network intrusion detection is a hot research topic in network security, in order to improve the accuracy of network intrusion detection, a network intrusion detection model (IPSO-SVM) is proposed based on improved particle swarm optimization algorithm and support vector machine to solve the problem of classifier's parameters optimization. Firstly, network intrusion detection rate is taken as the objective function, and support vector machine parameters are used as the constraint conditions to establish mathematical model, and secondly improved particle swarm optimization algorithm is used to find the optimal parameters, finally, support vector machine is used as classifier to build intrusion detection model, and KDD 1999 data is used to validate the performance in Matlab 2012. The results show that IPSO-SVM has solved the optimization problem of the classifier's parameters and improved detection rate, reduced false alarm rate, false negative rate of the network intrusion. Along with the increasingly important role of software in today's society, during the development process of software or system, testing becomes more and more important. With the wide application of object-oriented technology and the demand of automation testing, model-based testing (MBT) approach have been approved by software engineering. Although there are a number of publications discussing model-based software testing, we are lack of technical papers and publications presenting a review of the current advances in model-based software testing and automation tools. The goal of this paper is to review the model-based testing approach, it first discusses the popular models used in model-based testing, the test case generation methods, the testing target. Then, it summarizes MBT practices and discusses and compares the major model-based test tools. Finally, the paper close ends with a discussion of the challenges of current MBT approach, and where model-based testing fits in the future. Inspired by the deficiency of anonymous nodes, the descent of safety caused by the dynamics and the computation insufficiency of current trustworthiness mode, the paper proposes a computing mode of trust evaluation based on the dynamic Bayesian network. The mode calculates the trustworthiness from the direct trust and commendation one according to the historical interaction data. Time-effect factor and penalty factor are introduced in the paper to solve the problems of timeliness and malicious node. The efficiency and practicability of the mode are proved by the simulation experiment. This paper introduces a kind of basic functions of authentication server, high concurrency authentication server has the characteristics of concurrent two-factor authentication of the client and server, epoll is adopted in the system and the thread pool technology is adopted to ensure that its high concurrency, RSA and DES algorithm is used to guarantee the security of information transmission, the server has concurrent processing ability and further improves the performance of network security authentication server. Through the complete design example test, results show that the authentication server runs normally in the case of high load operating. The system has certain practical significance to process synchronously from the client request and certification research in a short time.