- EI核心期刊
An adaptive fuzzy controller based on fuzzy T-S model is designed to the question of bank-to-turn control system nonlinear model uncertainties and fast variable time characterizes. And a new sliding mode controller is designed by feedback linearization method. The T-S fuzzy model is used as a universal approximates to approximate the unknown nonlinear part of the control system. Also, Lyapunov stability theory is applied to guarantee the closed-loop stability. The asymptotic stability of nonlinear control system is proven, and the adaptive law is derived. Simulation demenstrates the controller achieved excellent tracking preformance.
A channel model for vehicle (bus) cabin is proposed based on frequency-domain channel measurements according to the ultra-wideband (UWB) frequency regulation specification of China. A quasi-Gaussian window with Gaussian transition band is used to extract the measurement signal specified by UWB frequency regulation of China in the post-processing of channel modeling. Furthermore, an automatic wavelet analysis based cluster identification algorithm is proposed for the time-domain measurements instead of visual inspection. The modified Saleh-Valenzuela model is exploited as the channel model and the channel parameters are extracted statistically from channel measurement data. Results is shown that the proposed vehicle cabin UWB channel model has similar excess delay, root mean square delay spread characterization and number of main paths with the measurement data.
The geometrical meaning of support vector machine（SVM）is analyzed. Integrated with the definition of health evaluation a method of network health evaluation based on SVM is introduced and translated into a classifying problem. The distance between the sample and health hyperplane is proposed to be the evaluation parameter. The health from the whole entity is evaluated and the subjective factor influences are reduced. The cloud model is used to realize the uncertained map between health degree and comment. The network simulation based on opnet shows feasibilities of this method.
A new congestion avoidance algorithm based on priority of path is proposed for multi-path wireless sensor networks. By monitoring the remaining space of the queue and congestion duration, the sensor node can update its congestion state index (CSI) in real-time. When CSI changes, the node calculates the its routing state index (RPSI), and notifies the neighbor nodes, which will update the records on the next hop nodes RPSI, and adjust the priority of the next hop path. The sensor nodes send different data through different routes based on the priority. Simulation shows that the proposed algorithm achieves satisfactory performance in reducing the occurrence of network congestion and the latency of network transmission.
A simple, flexible, and efficient method is proposed to generate hostbased attack graph. Based on this generating algorithm, the concept of ‘host-security-group’ is put forward, and according with its partitioning algorithm, an approach of analyzing network security by using theory of host-security-group is described. It is proved to be intuitive, efficient and accurate, as it describes the overall security situation of network and facilitates network security administrator to identify the key hosts.
Taking full advantage of function summary from interprocedural calls to get rid of the path explosion issue, a new Concolic testing method based on function summary is introduced. Function summary is generated by Concolic testing on interprocedural calls, it composes of the returned results and path conditions. Expansion test is replaced with function summary in the proposed method when dealing with the same interprocedural calls, that will prevent the repeated tests on interprocedural calls in traditional Concolic testing and thus improve the testing efficiency as well as mitigate the path explosion issue as the result of increased number of calls and growing depth of call stacks. Experiments of the prototype system indicates that the number of testing paths in the proposed method keeps linear growth along with increasement of the number of testing objects while the number of testing paths in the traditional Concolic testing grows exponentially.
A new cooperative spectrum sensing scheme in cognitive radio networks based on evolutionary game is proposed to analyze the secondary users’ transmission time and throughput, model the cooperative sensing as an evolutionary game, derive the dynamic procedure and optimal strategy for the secondary users, and develop a distributed learning algorithm as well. Analysis and simulation show that the proposed scheme guarantees sensing reliability without any performance loss and reduces the communication overhead significantly to achieve high throughput for secondary user in sensing cooperation.
For cell outage problem due to various faults in wireless access networks，an distributed time division-synchronous code division multiple access (TD-SCDMA) high load cell outage compensation algorithm is presented based on uplink target receiving power adjustment in open loop power control. In the presented algorithm, the parameters of cells neighboring failed cells can be automatically adjusted such that the coverage is maximized the given constrains on quality defined by cell-edge user throughput. Simulation shows that up to 78% of the user can be recovered at the cost of an acceptably reduced quality in the compensating cells.
Current work on fault detection wastes lots of probes to perform the detection process, and the work on fault localization has either high computation complexity or low diagnostic accuracy. To address the above problems, The probe selection for fault detection (PSFD) algorithm and integrated fault localization (IFL) algorithm are proposed. PSFD can detect the faults in the network with less probes and time, and IFL can perform fault localization much faster with higher accuracy. Performances of the two algorithms are validated by the experiment results.
A fault detection algorithm based on Dirichlet reputation for wireless sensor networks was proposed. The algorithm consists of three modules: measurement, reputation updating and fault judgment. The healthy level of sensor node in current time was evaluated in the measurement module. The work of maintaining and updating the reputation of the sensor node were performing by the reputation updating module. Fault judgment module determines whether a node is faulty or not, and whether the sensitive event happens or not. Simulation results show that the proposed algorithm has a high fault detection rate. Even 25 percent of nodes are faulty, 94.83 percent of fault nodes can be detected out and the false alarm rate is below 5%.
A dynamic programmingbased resource allocation algorithm is proposed to improve the performance of resource allocation in multi-user multiple-input multiple-output orthogonal frequency division multiplexing multicast system. The algorithm tries to find the optimal total capacity and the respective resource allocating strategy of the system by dynamic programming in a global manner after partitioning the available power into schedulable equal pieces. No repeated calculations for sub-problems is required, this algorithm has a good time characteristics. Simulations show that the difference of the total system capacity between the proposed algorithm and the optimal yet less effective algorithm is within 0.10%.
In order to eliminate the attack in a quick mode and ensure the safety of communication software servers without too much overhead on routers, with a three layer framework, a traceback approach with probabilistic packet marking Internet protocol is proposed based on cooperations, called cooperation based probabilistic packet marking (CPPM). The proposed algorithm adjusts the marking probability and marking information transmission fashion according to the load of routers in the network. It can reconstruct the attack path back to the intrusion source in a quick mode with less impact on the network. Simulations show the effectiveness and practicability of the algorithm.
The false-positive problems frequently occurred in applications of radio frequency identification (RFID) because of environment complexities. An efficient method is proposed to improve the RFID’s discrimination based on stability of signal strength. It can eliminate the interference between transponders by comparing their signal strength. Experiments indicate that this method provides a good method to improve RFIDs discrimination.
An overlay multicasting solution which combine a selected lightweight traditional multicasting routing algorithm and the related constrained model used to integrate multi-constraints is proposed. Based on the guidance above, a degree and quality of service related constrained fast lightweight multicasting routing algorithm is proposed. After analyzing the children-detect-and-father-response mechanism using both active and passive detections, the maintenance and control overhead is further reduced to deal with the node failure and the dynamic performance optimization, so a fast and lightweight multicasting protocol is achieved. Simulation for the multicasting routing algorithm proves the proposed.
The problem of analyzing the multiuser diversity (MUD) while taking both the gains of spatial diversity and spatial multiplexing into consideration is addressed within time division multiple access (TDMA) multiple-input multiple-output systems. A framework based on the concept of diversity-multiplexing tradeoff (DMT) is proposed. First, the multiuser outage probability is derived employing classical probability theory. Second, for any codes in TDMA system with given data rate, an upper bound of error probability can be derived. Then, for approximately universal codes in TDMA system with given data rate, a lower bound of error probability is obtained as well. The optimal DMT curve of a TDMA system is presented to reflect the MUD aimed at bridging the studies of MUD and DMT. Simulations coincident with theoretical analysis.
By using migration, virtual machines which located in different place can be consolidated. Then vain physical machines can be shutdown or hibernated in order to save power. Existing methods merely focus on migration cost or application-aware communication cost. Combined these two factors a virtual machines migration method is proposed based on migration cost and communication cost for power saving. It is shown that this method can efficiently save power in cloud and decrease communication cost.
As the structure of a tree augmented Naive Bayes(TAN)classifier is a directed acyclic graph, most of learning algorithms try to improve classifier's performance by selecting the directions of arcs. With scoring criterion of log conditional likelihood, however, it is proven that TAN classifiers with same links are equivalent after analyzing the searching space and equivalent classed of TAN classifier structures. Therefore, a learning algorithm without taking arc reversal into account is proposed. Experienments show that the proposed altorithm is effective.
A novel interference cancellation algorithm without requiring extra channel information, namely the information between user equipment and cooperated cells is proposed. By investigating the interference processed at the receiver in coordinated multipoint transmission single user multiple-input multiple-output (CoMP-SU-MIMO) transmission and the characteristics of dual polarized channel, combined with the existing limited feedback technologies, a novel interference cancellation algorithm was proposed. Utilizing the eigenvector’s orthogonality of polarized channel, the proposed algorithm could achieve better performance on interference cancellation, and the algorithm complexity can be reduced as well.
In order to solve the problem that the existing medium access control(MAC) layer protocols in wireless local area networks(WLAN) are less efficient, a fixed time and random extended implicit MAC Acknowledgment protocol is proposed on basis of extended implicit MAC acknowledgment (EIA). Since the proposed protocol brings in the leader and reduces the number of the multicast clear to send frames, the average channel holding time of transmitting data is reduced and the ratio of the real data transmission cost in entire multicast stage is increased. Simulations show that the proposed fixed time random extended implicit MAC acknwledgment protocol effectively improves the overall performance of multicast in WLAN and effectively increases the network goodput compared with EIA.
null
In order to reduce the adjacent spectrum leakage caused by the coherently non-linearity of radio frequency (RF) power amplifier（PA） in the communication system, a new multi-intervals polynomial model is proposed for solving non-linearity of the widespread doherty PA. According to system test, the proposed model reduces the adjacent channel leakage ratio (ACLR) up to 16 dB more than the polynomial model. Comparing to the digital predistortion solution of foreign company, the new model can further provide ACLR more than a 10 dB. Additionally, this model is more suitable for field programmable gate array (FPGA) implementation.
Buffered switches are analyzed firstly, and then a new scheduling algorithm based on unstable virtual output queue and unstable cross-point buffer is presented. The delay and stability performance are investigated carefully by simulations. It is shown that the new scheduling algorithm can offer very good performance on average delay and stability for no-uniform traffic with Bernoulli arrivals. The shortcoming of round robin-round robin algorithm is completely solved, and useful to design high performance combined input and crosspoint-queued (CICQ) switch.
After researching the load balancing problem in parallel rendering, an adaptive load balancing algorithm for volume rendering is proposed according to the character of volume data. This method is an improved Whiteman algorithm, it calculates the depth of four vertexes of grid to set the grid’s value and used time statistic to control the occasion of redistribution. Experiments show that this method can effectively improve the stability and capability of parallel volume rendering.
A fast live migration method of virtual machine is brought up for problem of long total migration time of live virtual machine migration in cloud data center. Through the formal definition and the performance analysis of the pre-copy model, a bitmap page which marked those frequently updated pages is added in the pre-copy approach. By judge in the iteration process, those frequently updated pages are put into the page bitmap, and transmitted in the last round of the iteration process only. A contrast between the pre-copy approach and the fast pre-copy approach is given. The fast pre-copy approach can distinctly reduce the total migration time of live migration, and achieve the purpose of fast live virtual machine migration.
The multiple-input multiple-output-orthogonal frequency division multiplexing (MIMO-OFDM) system can be impaired by channels estimation error over double selective (time varying and frequency selective) Rayleigh channels. The performance for MIMO-OFDM systems in presence of channels estimation error is investigated using zero-forcing detectors over double selective Rayleigh channels. Approximation to the distribution of complex Wishart distribution is used to derive probability density function of signal-to-interference-plus-noise ratio, average error probability and outage probability of the system. The analysis is shown good by numerical simulations.
Key words
A prediction algorithm based on user requests for prediction on user requests for social networking services (PUR-SNS) is proposed to improve the efficiency of web cache. The proposed algorithm is based on behavior pattern of user requests in SNS application and a forecasting mechanism is introduced. Simulations shows that PUR-SNS will improve the hit rate.
For color images, the state-of-the-art edge detection methods may result in loss of edges, imprecise edge positions, pseudo-edges or discontinuities. In order to solve these problems, an improved Canny edge detection algorithm is proposed. The algorithm combines photometric invariant differentials with color structure tensor to produce new image edge-gradient amplitudes. The canny detector based on new gradient amplitudes can detect edges of color images more precisely. Experiments demonstrate that the proposed method can well detect and classify variant color edges. The quality factor of edge detection evaluation shows that the color edge detection error rate has been reduced and the edge detection performance has been effectively improved
Based on analysis and the comparison of the existing platform as a service (PaaS), requirements of PaaS for internet applications are brought forward together with the conceptual model of PaaS. Under the guide of the conceptual model, the architecture of PaaS for Internet applications is proposed and some key technologies involved are introduced further. In present, the architecture of PaaS for internet applications has been applied in relevant project.
To improve the time-frequency concentration of the S-transform (ST) for the time-phase modulation (TPM), the generalized S-transform (GST) was adopted to the analysis of the TPM signal. Based on the window function of ST, a controllable parameter was added in the window function of GST. The window width for each frequency was optimized through the optimization of the added parameter and the time-frequency concentration was enhanced. The simulation shows that the GST approach can enhance the time-frequency concentration of the TPM signal in comparison to the standard ST, and the phase jump character of the TPM signal is retained which is helpful to the next processing of the signal detection.