An integrated method based on decision tree for Chinese pronominal coreference is proposed. The basic idea is to some extent that filtering out the negative examples based on rules and could compensate the drawback of decision tree that ignoring the relationship between attributes. The performance of the proposed method is tested on Chinese Treebank. In our experiments, the attributes and coreferences are manually labeled, and then the rule patterns are utilized to feature vectors following the decision tree of C4.5 algorithm. The success rate is 82.59%, in which the rate of personal pronouns and demonstrative pronouns are 87.60% and 75.21% respectively.
The downlink performance of orthogonal frequency division multiplexing (OFDM) frequency hopping system with time-frequency mapping patterns derived from Costas sequence was analyzed in a multi-cell multi-user scenario. Through the comparison of frequency hopping system and non frequency hopping system using system level simulations, the frequency hopping solution is proved to be effective against inter-cell interference. To mitigate inter-cell interference further, a new resource planning scheme based on “reuse partition” was proposed, it is suitable for frequency hopping This scheme provides interference avoidance by allocating a different prior slot for each cell. The resource in the non-prior slots also can be used due to the interference average induced by frequency hopping. It is demonstrated by simulation results that the proposed scheme will be able to mitigate inter-cell interference further when the system load is under a certain value.
Two high performance detectors were proposed. Both are sequential iterative detectors using linear minimum mean squared error (LMMSE) estimation. In the first detector, the finite alphabet property of modulated symbols was used. During the process of iteration, the prior information was also updated through posteriori probability calculations. In the second one, the decision feedback was used. Analysis and simulations demonstrated that compared to the existing interference cancellation method, and the proposed algorithms can effectively eliminate the error floor of space-frequency orthogonal frequency division multiplexing (OFDM) system over frequency selective multipath channels with affordable implementation complexity.
The availability assessment model of transport network was studied. An approach of running availability for assessing the running quality of transport network at runtime was proposed. By considering relationship of disabled circuit set and corpora circuit set, a runtime-oriented metric based on circuit services for running availability, service outage index (SOI), was defined. And through SOI, the quality of running status at assessing time is transformed into a number with the range of 0 to 100. And discrete SOI values of different time points compose a running availability curve. The network running status could be observed through this curve easily. Application results in a real network environment demonstrate the assessment model to be promising.
A video retrieval scheme based on global motion is proposed. A detailed global motion analysis for Moving Picture Export Group(MPEG)videos has been made and a camera motion model has been extracted. The model parameters are used as a motion features sequence, and the query by example of retrieval scheme is based on sequential match. Furthermore, according to global motion parameters motion analysis functions are constructed to segment exactly visual global motion and annotate semantic information, and then, the eXtensible Markup Language(XML) is used to describe the video global motion to realize the semantic query by Xquery engine. The experiments show that in the scenery videos, this scheme could query exactly and achieve semantic content retrieval to some extent, which has made an improvement in efficiency compared with other type of query schemes.
Two downlink packet data scheduling algorithms applied to orthogonal frequency division multiple access(OFDMA) broadband wireless access systems are proposed, which are multi-carrier proportional fair (MPF) scheduler and MPF algorithm enhancing capacity. The results demonstrate that they both provide fairness for users with explicit quality of service(QoS) guarantees, but MPF algorithm enhancing capacity achieves higher system throughput compared with MPF scheduler with more flexible parameter setting.
According to distributed multimedia application s’ characteristics, an entropy-based quality of service (QoS) Algorithm for grid resource is proposed for application’s QoS. The key idea of the algorithm is to construct the new metric-entropy and select the best-stability QoS with the help of entropy metric to improve the reliability and accuracy of the resource management. so as to provide QoS guarantee in the grid service environment whose changes continuously. The simulation results show that this works are effectively and reach near results.
In vector quantization(VQ),the initial codebook design is very important for VQ codebook performances. To overcome disadvantages of existing initial codebook algorithms, a new separating mean algorithm for learning vector quantization(LVQ)based upon self-organizing feature maps(SOM) was proposed. Experimental results for image VQ show that new initial codebook algorithm is better than random and splitting algorithm.
In order to reduce the complexity of the recursive lease M-estimate (RLM) algorithm, The impulse noise suppression function ??(•) is modified. The analysis proves that the modified algorithm still meets the linear equation. In addition, the method of estimating the key threshold ? and ? in function ??(•) is provided, and the modified RLM algorithm to estimating the impulse response vector in the time-varying fading channel is applied. The simulation results show that the modified RLM algorithm is more robust against consecutive impulse noise and have better convergence ability than conventional recursive lease square (RLS).
Adapting to the high variation environment in data grid, a data replica replacement algorithm based on value model is proposed and the arithmetic to calculate replica value is introduced. The algorithm is used on data replica replacement. When there is not enough storage space during replication, the replica whose value is the least will be deleted. This algorithm can improve and maintain a high overall throughput of grid jobs that access files in a data grid. Experimental results show that the proposed algorithm can obtain better performance for grid task to access data replica.
In order to minimize the number of electric equipment and wavelength in optical networks, a genetic algorithm of arranging the order of multicast traffics was proposed and therefore the traffics were groomed according to the orders. Two grooming algorithms based on non-restricted multi-hop provisioning were also proposed. The minimum value of required electric equipments of theory was calculated and compared with the result of genetic algorithm. Simulations show that the value of genetic algorithm is close to that of theory and genetic algorithm is applicable for traffic grooming under different conditions.
An improved threshold adjustment (ITA) algorithm is proposed, it can precisely guarantee the predefined frame error rate (FER) for pilot-assisted adaptive modulation and coding (AMC). An autoregressive (AR) model is introduced to improve the link prediction performance of ITA. Numerical and Link level simulation show that ITA can converge to the predefined FER, while Threshold Adjustment AMC algorithm proposed by 3GPP deviate the predefined FER. When feedback delay is large than 10ms. ITA algorithm with AR prediction could obtain higher throughput and more accurate FER performance than the one without AR prediction.
The cryptographic model of cheating immune secret sharing Function is introduced. By using injective functions, it proposes anew method for the construction of cheating-immune secret sharing which has the form (x,y)=x•π(y). It gives a sufficient and necessary condition for a quadratic Boolean function to be cheating-immune. It improves the conclusions given by Josef Pieprzyk
A static p-cycles deployment rule which every span is deployed by two p-cycles which are edge-joint just only with the span in optical network is proposed, so the dual span failures that can 100% survive is rigorously proven. The heuristic algorithms based on the rule are then suggested and the numerical results state that the p-cycles deployment with shared-p-cycles can decreases the protection capacity needed.
A method to numerically generate orthogonal ultra wide band (UWB) shaping pulse series is proposed based on wavelet and Hermitian matrix eigenvectors. In order to construct the Hermitian matrix, an appropriate pulse in less than 1 ns duration time was first generated based on wavelet. Then, the orthogonal pulse series are produced through Gram-Schmidt process using the Hermitian eigenvectors given by Hermitian matrix. The simulation results show that the power spectral density (PSD) distribution of the short time pulses generated with this method is constrained to the Federal Communications Commission (FCC) spectral mask for UWB, and the pulses have the characteristics of higher spectral utility rate, excellent autocorrelation and cross-correlation properties compared to the usual ones. This can meet the requirement for UWB to suppress the interference between multi-users.
A new algorithm to design ultra-wideband pulse was presented to maximize the spectrum utilization efficiencies. ultra-wideband (UWB)pulse can be obtained by windowing the ideal UWB pulse and convolution. The pulse designed here not only was limited to time so as to reduce the multiple access interference but enhanced the spectrum utilization efficiency under the spectral mask constraint, which improved the transmission quality of a UWB system. Furthermore, a flexible and effective method for designing pulses satisfying arbitrary UWB frequency emission mask is provided in this algorithm.
An adaptive congestion control algorithm in code division multiple access (CDMA) systems from the aspect of reducing cost is proposed. In this scheme, data rates in probabilistic manner according to a reverse activity bit (RAB) and an adaptive probability value broadcasted by the base station are increased or reduced in data mobiles, whereas the adaptive probability value is dynamically adjusted on the basis of the residual capacity of system. It is proved by analysis in theory and system level simulation that the adaptive congestion control algorithm can enlarge system throughput while controlling system load.
Aiming at the coherent disadvantage of transmitted reference (TR) receiver, a new TR-UWB (ultra-wideband ) receiver design scheme for low-data-rate wireless personal area network (WPAN) is proposed. The scheme not only takes into account noise suppression, but also considers the characteristics of signal transmission in dense multi-path environment. Computer simulation verifies the novel TR-UWB receiver can significantly improve the receiver performance.
Taking the example of designing classifier in intrusion detection system, the selection of training samples for classifier is studied. A new method is proposed for sample selection in large data set. First, it will reduce the size of selection problem via clustering, select samples according to the with-in cluster scatter value and coverage area of a sample. And it will retain boundary samples and discard most of the interior ones in each cluster. As reserving typical samples and reducing training samples, the generalization performance and training efficient of the classifier are guaranteed.
The security requirement for mobile agent-based electronic commerce transactions is summarized. An improved scheme based upon the strong non-designated proxy signature scheme is proposed. The new scheme adds new feature for customer privacy protection and effectively addresses security requirements when mobile agent moves between multiple optional hosts; it also can ensure the unforgeability of customer and shop’s identity, prevent reply attack, moreover, it will secret one shop’s identity and transaction information against others. The security analysis of the proposed scheme show that it meet all security requirements summarized in section one.
The architecture on the mobile P2P (peer-to-peer) system which is composed of cellular Ad hoc network and 3G core network is proposed. The construction solution and the functional architecture of each network entity are described in detail, and at the same time, the communication processes between peers are analyzed. A simple and efficient index server searching mechanism is presented in the constructed Mobile P2P system. Simulation shows that, compared with flooding searching mechanism, the energy consumption in mobile P2P system will be improved when in use of index server searching mechanism, to some extent, the deficiency of mobile termination is remedied.
To improve the efficiency and scalability of conflict detection for multi-dimensional classifiers, a new algorithm, based on grid of trie (GoT) algorithm, was proposed. The new algorithm uses Patricia trie, constricts the length of Internet protocol (IP) prefix in order to use Hashing technology, and improves the performance of the algorithm by adding ingress and egress of firewall for each filter.
Constructing the video summarization in semantic level is very important. An algorithm of video summarization based on support vector machine (SVM) in semantic level for the natural scenery documentary is presented. The shot key frames are classified by SVM and the frames constructing the video summarization are selected from the shot key frames of every class by the importance function introduced. The scalable video summarization to fine level of detail can be achieved by changing the threshold of important function. Experiment results indicate that the proposed algorithm performs satisfactorily.
Timing synchronization algorithm based on training sequence in multi-input multi-output-orthogonal frequency division multiplexing ( MIMO-OFDM) system were demonstrated. With a view to hardware implementation of timing synchronization algorithm in MIMO-OFDM system, the complexity of timing synchronization algorithm implementation with FPGA(field programmable gate array) was analyzed according to multiplying and adding frequency in unit time. To make the timing synchronization algorithm in MIMO-OFDM system less complicated during FPGA implementation, after study on sub-modules in individual timing synchronizing module and relationship among multiple timing synchronizing modules, a simplified implementation scheme for MIMO-OFDM system timing synchronization algorithm was brought forward, and statistics on resource occupation of such a scheme in VirtexII-Pro series FPGA of Xilinx Company was given. Study shows the simplified implementation scheme is applicable in hardware implementation of MIMO-OFDM system timing synchronization algorithm.
Based on improvement of authentication method of typical proxy signature scheme, the secret share was distributed by a proxy group administrator。A threshold proxy signature scheme with known signers was proposed. The scheme has the non-repudiation property. The non-repudiation property provides the scheme the ability to identify the actual signers of the proxy signature, and then to avoid the abuse of signatures. Furthermore, the proposed scheme can withstand conspiracy attack.
A simplified and optimized low density parity check (LDPC) codes for orthogonal frequency division multiplexing (OFDM) transmission is proposed. Making consideration of the channel fading characteristic and iterative decoding principle, high variable degree is used in deep fading sub-channel and low variable degree in light fading sub-channel. Simulation results show that the scheme exhibits a gain of 3 dB at bit error rate (BER) of 10-5 when code length is 1024 and code rate is 1/2 compared to regular codes and reduces the complexity of optimization design.
A distributed nodes configuration protocol for virtual backbone-based mobile ad hoc networks(MANETs) is presented. The prophet allocation algorithm is executed in backbone designedly, the approach could be applied to large scale MANETs with low complexity, low communication overhead, and low latency. Finally, the performance of the solution is evaluated through simulation experiments.
Multi-user signal synchronization and separation for the interleaved orthogonal frequency division multiplexing (OFDMA) uplink are investigated. In interleaved OFDMA uplink, each user’s signal has a special periodic structure within an OFDMA block. Firstly, by utilizing the signal structure characteristic and subspace-based method, multi-users’ carrier frequency offsets can be obtained. Next, minimum mean square error (MMSE) filters are applied to the received signal to recover single user signal wave thus implementing user separation and frequency synchronization at the same time. Finally, based on the recovered single user training signal, each user’ timing offset can be estimated. The generalized chirp-like ( GCL ) sequence has the ideal periodic auto-correlation function, optimum periodic cross-correlation function and low peak-average power ratio (PAPR). Therefore, GCL sequences are used as training sequences to estimate each user’ timing offset. Simulation results show the proposed scheme is effective in multi-path fading channels.
CHEN Wei1, YANG Yi-xian1, NIU Xin-xin2
On the basis of introduction of A5/1 linear initiation weakness, by the characteristic of non-balance output in inter-clocking mechanism based on majority logic is analyzed, and the correlation attack weakness is discussed. Meanwhile, that thrown-out of original 100 bits output in A5/1 algorithm that can not approve its security is pointed out. For such weakness, inter-clocking mechanism is extended to the originate process of A5/1 to eliminate such weakness. It is concluded that A5/1 algorithm is safe under such attacks.
Security is a challenge in wireless sensor networks. The potential security threats, the key issues and design principles for achieving the security of wireless sensor network are analyzed. Integrated secure system architecture about sensor node to secure sensor network is presented. As a conclusion, the security requirements, such as confidentiality, integrity, authentication, freshness are provided while consuming little system resource and the impact on the whole network’s security are minimized after some sensor nodes are captured and compromised as well.