73 results about "Multiple attack" patented technology

The embodiments of the present invention disclose an event detection method and device. The method includes: predefining event-based detection rules with a predicative context-free grammar; generating by parsing the detection rules a parsing table of pushdown automaton which supports parallel parsing; receiving an event to be detected; and analyzing by a controller the event to be detected according to the parsing table, to obtain a detection result. The present invention is especially applicable to detection of network attack events. The embodiments of the present invention detect the attacks with a predicative context-free grammar on the basis of events, and ensure a close combination of a protocol parsing process and an attack detection process, as well as a close combination of multiple attack detection rules, thus decreasing unnecessary calculations. In addition, with an optimized parallel pushdown automaton, the embodiments of the present invention can efficiently analyze the predicative context-free grammar. Consequently, besides hierarchical processing capability and state description capability, the embodiments of the present invention deliver high efficiency.

The invention discloses an image watermark embedding and extracting method based on a back propagation (BP) neural network compressed domain. The image watermark embedding and extraction method is based on Arnold transformation, to support BP neural network image compression processing, and the watermark is embedded into the neural network hidden layer to the output layer weight of the carrier image almost no effect, invisibility is very strong, and more secure. In the compressed domain that BP neural network, the carrier watermarked image even after multiple attack, is still able to remember the local relationship between pixels in order to achieve the correct detection of the watermark signal, making this embedding and extracting methods images of a variety of conventional attack has strong robustness, good balance contradictory and robust image watermarking imperceptibility between, and enables the blind to detect the watermark image.

The present invention describes a method for simulating a hacking attack on a Network, wherein the Network comprises at least one of a plurality of data processing units (DPUs), a plurality of users and a plurality of communication links, to assess vulnerabilities of the Network. The method includes receiving one or more scan parameters for the Network. Further, the method includes creating at least one master agent by a system to gather information about the Network, wherein the information pertains to critical and non-critical information about the Network. The method includes creating an Information Model and then incrementally updating the Information Model during the hacking attack. The Information Model is the abstract representation of information collected by the system. Furthermore, the method includes generating a Multiple Attack Vector (MAV) graph based on one or more scan parameters and the Information Model. MAV has the ability to combine plurality of low and medium severity vulnerabilities associated with the data processing units (DPUs), users and communication links, correlate vulnerabilities in combination with Information Model and generate high severity attack paths that can lead to compromise of the Network. Moreover, the method includes launching one or more attacks based on the MAV graph to compromise the Network. The method further includes installing atleast one slave agent on the compromised Network to perform the one or more attacks in a distributed manner. Moreover, the method includes performing a multi stage attack by using the at least one slave agent and the at least one master agent by repeating above steps. Finally, the method includes generating a report by the scan controller, wherein the report contains details about the compromisedNetwork and the vulnerabilities of the Network.

The embodiments of the present invention disclose an event detection method and device. The method includes: predefining event-based detection rules with a predicative context-free grammar; generating by parsing the detection rules a parsing table of pushdown automaton which supports parallel parsing; receiving an event to be detected; and analyzing by a controller the event to be detected according to the parsing table, to obtain a detection result. The present invention is especially applicable to detection of network attack events. The embodiments of the present invention detect the attacks with a predicative context-free grammar on the basis of events, and ensure a close combination of a protocol parsing process and an attack detection process, as well as a close combination of multiple attack detection rules, thus decreasing unnecessary calculations. In addition, with an optimized parallel pushdown automaton, the embodiments of the present invention can efficiently analyze the predicative context-free grammar. Consequently, besides hierarchical processing capability and state description capability, the embodiments of the present invention deliver high efficiency.

The invention relates to a device and a method for testing the structural fatigue of an H-shaped vertical shaft wind turbine blade. The device comprises a support base component, a distribution beam, a loading hoop, a support connecting rod and an eccentric motor, wherein the distribution beam, the loading hoop and the support connecting rod are connected together by using bolts so as to form a whole load distribution system, so that the fatigue load which acts in an upper and lower reciprocation manner can be effectively conducted. When the device is used for experiment, the fatigue bending moment distribution of the blade under various working conditions can be simulated by adjusting the length L of the distribution beam and the position B of the loading hoop, and the fatigue stress state of the blade in axial torsion and multiple attack angles can be simulated by adjusting the structural modes of the support base component and the loading hoop; the strain distribution of the blade under the fatigue loaded condition can be measured, and furthermore, the fatigue structural property of the blade can be evaluated; by only exerting one external fatigue load to the device, combined fatigue loading modes such as bending, torsion and multiple attack angles of the blade can be achieved, and the device is simple and convenient to operate, accurate and reliable in result and relatively high in universality.

The invention discloses a power system vulnerability evaluation method based on multiple attack strategies. The method comprises the following steps: 1) constructing a power grid model; 2) evaluatingthe importance of nodes in the network from four aspects of node degree, betweenness, local centrality and node fault loss, and completing the formulation of an attack strategy according to the importance of the network nodes; 3) establishing a betweenness virtual flow model to simulate a cascade fault propagation process in a power system; and 4) establishing a vulnerability index based on the betweenness virtual flow model, simulating a network attack by adopting a plurality of attack strategies, and evaluating the vulnerability of the IEEE300 power system network according to the change ofthe vulnerability index. Aiming at the vulnerability of the power network under the condition of various attack strategies, the betweenness virtual flow model and the various attack strategies are established to analyze the performance change of the power grid under the cascade failure. The importance of nodes in a power network can be identified, and decision support is provided for maintenance work of power system managers and power system fault prevention of designers.

The embodiment of the invention discloses a method for acquiring multiple pieces of attack behavior information. The method comprises the following steps: receiving at least one piece of attacker information, attacker flow information and corresponding attacker identification information created according to the attacker information; forwarding the attacker flow information to the honeypot throughthe first port; sending the mapping relationship between the attacker identification information and the first port to the honeypot through the second port and temporarily storing the relationship; and enabling a honeypot to send the flow analysis result to the system service module according to the temporarily stored mapping relationship between the attacker identification information and the first port. According to the technical scheme disclosed by the invention, the effect of completely recording and restoring the attack timeline under the condition of multi-node management can be achieved.

The invention provides a digital fingerprint-based PDF document copyright protection method. The method comprises the steps of firstly, designing encryption rules; secondly, hiding digital fingerprints through the information embedding algorithm to obtain a cipher-carrying PDF file; thirdly, extracting the digital fingerprints through the information extraction algorithm to convert the cipher-carrying PDF file to a plaintext. The method can track the identity information of a document leaking person and provide a new layer of protection aside from traditional file protection; the method uses a mixed method to hide identity information and increases the redundancy of hidden information through cyclic embedding, thereby improving the robustness of hidden information, being capable of resisting multiple attack modes and guaranteeing high security of hidden information. The method has certain practical value and a certain market promotion prospect, and not only can be used for protecting secret documents of enterprises and offices but also can be widely applied to copyright protection.

The invention relates to a three-dimensional mesh model blind watermarking method based on image discrete cosine transformation. The three-dimensional mesh model watermarking method involves selecting two stable geometrical characteristics to facilitate ordering to obtain a unique three-dimensional vertex sequence and adding watermark data to the texture coordinates of a vertex so that the vertex sequence with a watermark embedded therein can also be correctly obtained even after affine transformation, and the watermark can be correctly extracted accordingly. By using the method provided by the invention, original grids and pretreatment are unnecessary for watermark extraction. Multiple attacks can be resisted, including rotation, zooming, translation, noise and vertex reordering. At the same time, an algorithm also has a quite large watermarking capacity because when a carrier image is taken as the watermark for adding to a three-dimensional mode, the watermark transparency of the carrier image does not have to be considered, and thus the watermarking capacity is greatly improved.

The invention discloses a privacy protection method and device for inquiring an average score, a computer device and a storage medium, wherein the method comprises the following steps: calculating a personal privacy budget threshold; generating query conditions and judging whether the current privacy budget of a user exceeds the personal privacy budget threshold; calculating the sensitivity of thesystem if the privacy budget threshold is not exceeded; according to the sensitivity of the system, conducting noise processing on the real score; conducting numerical processing on the real score after noise processing, and sending the numerically processed score information to the user. The privacy budget threshold is calculated, and when the current privacy budget does not exceed the preset privacy threshold, the sensitivity of the system is calculated, noise processing is performed on the real score, and then the real score is sent to the user after numerical processing, so the security of the score inquiry system is improved in the face of multiple attacks at the same time organized by multiple people.

The invention provides a preparation method of a composite ceramic. The preparation method of the composite ceramic comprises the following steps that fiber cloth is bonded on the surface of a bulletproof back plate; a plurality of packaged ceramic small blocks are bonded on the surface of the fiber cloth; the fiber cloth is bonded on the surface of the plurality of packaged ceramic small blocks to form a ceramic plate; and the ceramic plate is cured to obtain the composite ceramic. The composite ceramic prepared by the method provided by the invention forms a new integral ceramic plate, and has the shock wave isolation effect, so that the anti-multiple attack ability is remarkably enhanced; the ceramic small blocks are bonded to one another, so that the overall bending strength is remarkably improved, the structural integrity and the bulletproof performance of the ceramic small blocks are not reduced under a long-term static load or dynamic load; and after impact, the damaged ceramicsmall blocks can be removed and replaced with the new ceramic small blocks in order to be used for multiple times and improve the use efficiency of the ceramic plate, and even in an emergency situation, the ceramic plate can be repaired. The invention further provides a composite ceramic.

The invention discloses a vector space data digital fingerprint method based on GD-PBIBD coding. The method comprises the following steps: constructing a fingerprint code by using GD-PBIBD with limited conditions, scrambling the fingerprint sequence to be embedded by using Logistic mapping; extracting feature points of vector space data through a D-P algorithm, performing DFT on the feature pointsto obtain a phase coefficient and an amplitude coefficient, then applying a QIM method to embed a fingerprint into the amplitude coefficient of a DFT domain, and finally applying inverse DFT to obtain the vector space data containing the fingerprint. The colluder can be tracked according to the Hamming distance between the suspicious data fingerprint and the original data fingerprint. The methodis simple in coding construction, can accommodate more users compared with a BIBD code under the condition that the code length is fixed, and is high in coding efficiency. Robustness to single-user fingerprint attacks and multiple attacks is good, and the problem of data copyright recognition can be solved. Multi-user minimum value and maximum and minimum value attacks can also be resisted, all traitors can be tracked, and a basis is provided for vector space data traitor tracking.

The invention discloses a programmable controller for a game peripheral. The programmable controller is characterized by comprising a micro-processor, a key input interface, a PS2 keyboard interface and a USB keyboard interface, wherein the PS2 keyboard interface and the USB keyboard interface are connected with a PC (personal computer) through an interface protocol, and the micro-processor only recognizes one keyboard interface during working; the key input interface comprises two setting keys, four direction keys, two auxiliary keys, multiple attack keys and one reset key, and the keys except the setting keys and the reset key all correspond to a key value of a universal key of the PC; multiple conventional game key layouts are stored in the micro-processor, and factory default key value parameters are restored by pressing the reset key for a long time. The programmable controller has the advantages that the programmable controller is no longer limited to several unchanged arcade games supported by gamepads and can be widely applied to various online games and single games. The games can be configured in the micro-processor before delivery and can also be automatically downloaded through software configured on the PC.

The invention discloses a network alarm information clustering method based on attribute correlation. The method comprises the following specific steps: arranging network alarm information; clustering the network alarm information with the same attack source or the same attack target according to the quintuple features; clustering the network alarm information with the same data characteristics and attack types according to the attack type characteristics of the alarm information to obtain a network alarm information set with the network alarm information type characteristics; clustering network alarm information based on attack occurrence time; and clustering network alarm information and fusing again. The alarm information clustering of the space-time global perspective is realized based on the multi-dimensional attribute feature clustering, which is matched with the situation that an attacker often carries out combined attack on an attacked target by adopting multiple attack technical and tactical methods in an actual attack event, and can cover the full view and important details of the attack event; and better basic conditions are prepared for attack event analysis and attack scene restoration.

The invention provides a copyright protection method of image data. Character information comprising copyright information is encoded to an image sequence formed by combining a limited number of binary image templates, and through combination with image feature points and a wavelet transformation technology, the binary image sequence is embedded into an image. For the purpose of fending off rotation and zoom attacks, preset template linear points are embedded into a circular ring Fourier transformation area of SIFT feature points of the image. During extraction, through analyzing distribution of the template linear points, rotation and zoom deformation parameters applied to the image are estimated, the image is restored by use of the parameters, and then the extraction is performed, through matching watermark binary images obtained through the extraction with preset template images, a watermark character sequence is restored according to an encoding algorithm. The algorithm has quite high robustness and solves the problem of incapability of simultaneously fending off multiple attacks especially cutting and rotation zoom geometric attacks by use of a conventional image algorithm.