This paper presents a method to recovery high-level control structures from binary executables, which is based on propositional calculus. This method formalizes the control dependence relation between low-level instructions to propositional variables. These variables can be propagated and calculated follow the execution path of the object code, and high-level control structures can be identified by some specified results of propositional invariants. Experimental results show this method can recognize and recovery loops and conditionals effectively, and have the ability of analyzing the predicated instructions.
On the basis of the introduction of the fountain codes and the Luby Transform(LT) codes, the Raptor codes are explored. Then the improving Raptor codes based on RS codes through the analysis of the precoding method are put forward, which can overcome the deficiency that time and space overhead is not fixed in the LT code. The improving method has many advantages such as high decoding rate and provides appropriate solution to the problem of safe transmissions and high availability in the Internet transformation. Simulation indicates that the improving Raptor codes can improve correctly decoding probability.
In order to predict the surrounding ground settlement of deep foundation pit excavation accuracy, a new method GGEP(Grey Gene Expression Programming) is proposed to predict the surrounding ground settlement of deep foundation pit excavation, which combines Gene Expression Programming(GEP) with Gray Model(GM). The new model is verified by Shanghai Metro engineering. Experimental results show it is higher goodness of fit with actual monitor data than traditional method. Not only its forecasting accuracy is higher, but also it can self-study to adapt new environment.
AES and Camellia are the two most widely used block cipher algorithms. The reconstruction and high-speed implementation characteristics of them are very important in both theories and practices. In this paper, on the foundation of analyzing the algorithms, the reconfiguration of AES and Camellia is studied. Based on the S-box transformation, reconfigurable architecture of the two algorithms is given by parallel processing and reconfigurable technology. Moreover, high-speed realizations of AES and Camellia algorithms are proposed using the reconfigurable architecture. Testing results show the scheme has a higher speed and it can reduce the area cost greatly.
Some achievements in the field are emphasized such as database technology, CAD systems development technology, network technology and three-dimensional visualization technology and so on. Some relevant technologies are developed such as electronic measurement traverse, mineralize model CAD management system, graphics system and information management system of geology and survey. The geological database, hydro-geological database, three-dimensional deposit model are established. FTP server of geology, survey and mining are set up. Three large stopes are monitored using cavity monitoring system and three-dimensional visualization of stope model is established. All kinds of mining index of stope is calculated accurately. The research indicates that the widespread application of geology and survey engineering information technology have extremely vital significance to speeds up our country numeral mine construction.
This paper introduces a Tree-Augmented Na?ve Bayes(TAN) text categorization model, analyzes its problem of threshold selection, and proposes an Automatic TAN(ATAN) text categorization framework. Two algorithms based on ATAN are compared to the BL-TAN with the best classification performance at a specified threshold both on Chinese and English imbalanced datasets. Results show that algorithms based on ATAN have higher performance than BL-TAN.
Frequent database connection generally influences the performance of large-scale Air Traffic Control(ATC) system. In order to solve the problem, this paper improves OCI9 database call interface by using connection pool technology, so that the system costs brought by the frequent creating and destroying of database connection object are avoided and the performance of the system is enhanced. Experimental result proves the effectiveness of the method, and a stack overflow error of Oracle InstantClient Version 10.2.0.5 is found.
Traditional mining method is easy to omit rules which is comprised by high frequent item set and it is non-standard for function degree value range. Aiming at these problem, this paper presents a positive and negative association rule mining method based on comparison influence degree. It can control the quantized result between [-1, 1], effectively mine positive and negative association rule. Application of this method into esophagus cancer case data which is a high incidence in Chaohu of Anhui, it can mine some potential relevant factors and give some corresponding measures to guard against esophagus cancer.
Locating main function from binary code is not an easy task when there is no symbol table in it. Aiming at this problem, this paper presents a novel approach for Portable Executables(PE). It analyzes the execution process of a PE from the entry point to abstract some patterns for locating. Instruction pattern matching is used to locate the address of function main() from a PE file. Experimental result shows that the approach based on pattern matching is favorable and effective on locating function main() from portable executables in different platform. It can improve the analyzing ability of binary code analysis tools.
Aiming at the defects of low effectiveness and function missing of traditional file test, this paper proposes a file fuzzy test method based on file format description. It gives file deduced rules which show the relationship between structure variables based on file format description, generates different kinds of files according to one special file format and performs fuzzing tests on them. Experimental results show that software vulnerability can be found by the method effectively.
Aiming at redundancy computing problem of spatial association rule mining algorithm, this paper proposes a spatial association rule mining algorithm based on numerical index, which is suitable for mining monolayer transverse spatial association rules. It uses ordering to generate numerical index in the course of bottom-up and top-down search, and uses numerical index to prune repeated candidate and delete redundancy computing in the course of double mining, so that the efficiency of the algorithm is improved. Experimental result indicates that the algorithm is more efficient and effective than other algorithms.
For problems of current backup system, such as large amount of duplicate data, too many data transmit over the network, and low data security, this paper presents a new file level backup system, which is built on the open source Bacula backup system. It uses the disk as a storage medium, based on content-based addresses to store data, and carries out de-duplication technology to save storage space. Performance test results show that the system can effectively ease the pressure on storage and network transmission, while improving data security.
Aiming at the problem of filtering malicious information which contains transformed keyword, this paper presents a feature extraction method in Chinese text based on flexible matching. The method adopts flexible matching technology to identify transformed keyword, improves the computational method of feature term weight in Vector Space Model(VSM). The keyword which has transmutative form is endowed high weight to enhance extraction efficiency for feature information. Experimental result shows that the method of feature information extraction for filtering has high recall in the condition of ensuring precision.
This paper presents a network performance comprehensive evaluation approach to the performance requirements of Wireless Sensor Network(WSN) in industrial field. The indexes weights and the comprehensive evaluation value are calculated by objective and subjective combined method and the linear weighted method respectively. Experimental results illustrate effectiveness of the proposed method.
In Wireless Sensor Networks(WSN) MAC addressing technologies currently, the schemes using Huffman code short the average address length while in need of several negotiation. The schemes based on cluster structure of networks have clear frames but with unfair address assignment, and that based on geographical location information can provide assistant to routing while the precise location information is hard to achieve. Based on the analysis of address reuse principle, this paper enumerates several representative schemes and compares their performance diversities and application area.
In order to solve fault-tolerant problem in Network on Chip(NoC), this paper proposes a low-latency and reliable delivery algorithm based on end-to-end communication. By sending redundant data packets initiatively from sending end and dividing data into packages to increase the utilization of link, the algorithm can reduce delay. The sending end sends next package only after receiving the affirmation of current package. Constant sending data and end-to-end feedback promise the correctness of transmission, which guarantee high reliability. Experiments on NS-2 prove that the algorithm has low delay and high reliability of communication on a chip, and can handle transient errors in transmission efficiently.
Based on the distributed wireless networks, this paper proposes a new MAC mechanism protocol——Distributed Dynamic Frame Reservation Multiple Access(DDFRMA) which supports voice and data transmitting together. The Markov model is built to analyze the relationships between nodes and the voice packets dropping probability. Simulation results show that the protocol can support voice transmitting of distributed networks, improves the voice capacity of access mechanism, compared with the CSMA/CA mechanism.
Owing to the particularity of the wireless sensor network, the energy supply becomes a major technical problem that constraints its constant development. However, rectenna which is able to carry out the wireless power transmission is exactly in line with the technical features of the wireless sensor network. This paper designs a new type of rectenna. It employs a method of directly matching a rectifying circuit to a microstrip-patch antenna, as a result, the rectenna’s efficiency is improved and its dimension is reduced. This design realizes conversion efficiency of 82.4% at 5.8 GHz, so that it can satisfy the need of wireless sensor network better.
This paper proposes a hop-by-hop cross-layer congestion control algorithm, according to detecting buffer occupancy ratio and congestion degree. This algorithm adaptively adjusts channel access priorities in MAC layer based on open loop hop-by-hop rate control. It can adaptively adjust the information sending rate according to local congestion condition of network nodes. NS simulation results show that the algorithm can improve network performance and congestion control efficiency.
Two forgery attacks are presented on two classes of blind proxy multi-signature schemes which have the properties of blind proxy signature schemes and proxy multi-signature schemes. In order to resist the two attacks, two classes of improved blind proxy multi-signature schemes are proposed. Result is proved that improved blind proxy multi-signature schemes not only resist the above two attacks, but also are more efficient than others. The computation cost is less and the length of proxy signature is shorter.
The time-memory trade-off algorithm is a method for quickly inverting a one-way function, using pre-computed tables. This paper combines time-memory trade-off algorithm with Word document decryption, based on rainbow table, gives a definition of a one-way function according to the encryption mechanism of Word document, a time-memory trade-off algorithm suitable for Word document is designed. Experimental result shows that the algorithm has high performance on attacking Word document.
The false positive rate of the HDPM scheme can be much higher than it is claimed when the number of ingress router interfaces is larger than the number of attackers. This paper proposes a Novel MAC-based Authenticated Deterministic Packer Marking(NADPM) scheme for IP trace. This method uses IP address and MAC authentication information based on different network protocols to choose for different packet marking the median. The implementation and evaluation demonstrates NADPM algorithm compared with other HDPM algorithms, the false positive rate reduces a lot, and can trace the maximum number of simultaneous attackers increasing to 140 000.
The exploiting of the software vulnerabilities is a hotspot of information security research. The shortage of current vulnerability exploiting implements is analyzed. The function demands of vulnerability exploiting and analysis based on Fuzzing are emphasized. A heuristic construction of abnormality data based on structural storage characteristic is proposed. The file format vulnerability intelligent exploiting and analysis system are designed and developed. The system’s software structure and run mechanism and critical technology are given. Experimental result proves its effectiveness and intelligence.
Aiming at the shortcoming of the traditional network base stations planning method, a new method is proposed based on immune algorithm. The mathematical model of multi-objective optimization problem for base station planning is given. The immune algorithm uses the mechanism for regulating the concentration of selection probability, the exclusion algorithm of neighboring memory cells, loop crossover and mutation operations. All ensure the diversity of the antibody and the Pareto optimal solution set of the distribution. Experimental results show that the algorithm can effectively find the optimal distribution scheme of base station and the coverage reaches to 97.6%.
To solve serious interference problems of redundant edges for edge-based statistical features of license plate location, a new vehicle license plate location method based on Pulsed Coupled Neural Network(PCNN) is proposed, which introduces the simplified PCNN model based on traditional license plate location algorithms. In the algorithm only the candidate area is produced by PCNN iteration in order to reduce the computational complexity and improve the rate of license plate location. Simulation experiments with 300 frames of vehicle license plate image that are taken under various kinds of conditions can get a 98.3% extraction rate.
A new algorithm is proposed for small objects tracking, which uses both Gaussian Mixture Model(GMM) and Particle Filter(PF). Considering to the problem of missing in detecting on small objects, the algorithm uses the information of objects feed back by PF to instruct the GMM in background modeling. The results of experiment show that the algorithm can achieve auto-detecting and auto-tracking of multi-objects while overcoming the missing of small objects in former algorithms.
Aiming at the problem that grid deformation algorithm needs fix point and operating point defined by user under multi-fix point restraining, this paper presents a quadratic interpolation deformation algorithm of polygon central point vector. According to the moving area of rotation between source and destination polygon central point vector and the difference between the corresponding vectors to build a similarity function. It adopts quadratic Bezier interpolation method in the process of deformation and uses improved dynamic programming algorithm in the process of corresponding. Experimental results show that this algorithm can reduce skew angle of polygon in the process of deformation, and it has low computation complexity, short corresponding time, natural deformation effect.
Aiming at the problem that the accuracy is low when Fuzzy C-Means(FCM) clustering algorithm segments MR images with low signal-to-noise ratio, this paper proposes a modified algorithm named MS-FCM, and applies it in MR image segmentation. Considering the property that the membership values corresponding to the neighboring pixels which belong to the same cluster are similar, it filters membership data sets in the iterate process to decrease the influence on clustering accuracy caused by noise. Experiments on simulation brain MR images with different level noises and real brain MR image show that MS-FCM can improve the accuracy of segmentation.
Aiming at the large computing costs problem for gradient orientation of the image applied in rotation-invariant template matching, this paper proposes an improved rotation-invariant template matching method based on gradient orientation code. It uses integral histogram to reduce the computing costs of computer histograms, and uses pixel-jump method to reduce a lot of ineffective matching operations. Experimental result shows that the method can speed up from 3 to 6 times in the premise of ensuring matching accuracy, so it can meet real-time requirement.
To meet the demands of application situation of mass data throughput, this paper researches memorizing control algorithm which supports high treatment efficiency and data continuous read/write operation, and uses Field Programmable Gate Array(FPGA) to design SDRAM embedded control algorithm. The controller uses CMD order mode, assigns time-delay of continuous read/write based on burst length, enhances the data processing efficiency by team working between data channel control and read/write operation. Testing results show that running frequency of this controller is higher than 100 MHz, data processing efficiency is above 95%, and it applies to data buffer of video acquisition and large size display control of LED.
For satisfying the urgent need of enhancing the operation reliability and cutting down the malfunction rate by broadly applying lots of embedded control nodes based on the field-bus standard in the distributed control system, a kind of embedded control node according with the Modbus communication protocol is developed. In this industry node, an AT89C51 is selected to work as the main controller, the hardware design and work mechanism about some function modules are discussed, such as Modbus communication module, monitor and control module, display module etc., and the Modbus communication protocol and communication flow are given out, so that a supervisory control station can effectively monitor and control the work state of all nodes through this Modbus communication interface.
This paper proposes and implements an efficient multi-processor Network Access Controller(NAC) based on message transfers in Network-on-Chip(NoC). It supports receiving and transmitting simultaneously, contains a series of configurable registers, and introduces a Message Table(MT) based receiving method which makes data transfer more efficient by recording and updating receiving configuration information of messages automatically. The hardware scale of the NAC is about 20 443 gates and it operates at 300 MHz AHB clock in SMIC 0.18 μm technology.
In order to achieve enterprise application interaction and integration, this paper introduces the concept of enterprise resource, builds domain ontology resource library by identifying and describing enterprise resource using OWL. Enterprise business functions are encapsulated as Web services, which are described under the framework of this model, and mapping between services and domain ontology is established. The enterprise resource service encapsulation model is realized with OWL-S, and the correctness and feasibility of the model are proved through prototype system.
This paper introduces the structure of the personalized recommendation system and the design method. It states the process of establishing user interest model, including data collection and pretreatment, data refining, fuzzy linguistic processing, defuzzification, and satisfaction calculation. It analyses clustering items and processes, implements self-adaptive recommendation services based on the fuzzy adaptive resonance theory. Experimental results show that the system is stable and effective to predict users interest preferences and capture users excursion of interests.
The time difference between the local 1PPS time reference and the satellite one including outliers is important effect on instability of the satellite 1PPS signal and ambient interference and measurement noise in GPS/BD “dual-mode” satellite common view system. An adaptive Kalman filtering based on residual chi-square test is presented, which is effectively resistant to outliers and increases the result of common view comparison.
This paper finds the program’s bottleneck in inefficiency of loop’s operation by analyzing the traditional Fourier transform, besides, when the program is running, it is only distributed in one hardware core, so it doesn’t make the best of the multi-core. Aiming at the above problem, through researching the Intel Threading Building Blocks(TBB), the paper puts the loop’s operation into several mutually independent partition, and puts these partitions into multi-core, so it implements the traditional algorithm parallelization, the parallel program achieves an excellent effect.