- EI核心期刊
Short、open、load、through (SOLT) and through、reflect、 line (TRL) calibration techniques were described. The 12-term system errors for two-port calibration were shown, and the actual S-parameters device was derived from the measured S-parameters by using matlab programs. Comparison between the matlab programs results and the vector network analyzer measurement data validated the correctness of the calibration approaches and procedures. The SOLT and TRL microstrip calibration standards were designed for microstrip filter measurement.
Strong barrier path construction for directional sensor networks was focused on by optimizing the sensing orientations of directional sensors. The intersection point set and associated intersection point set were defined to describe the relationships between directional sensors and region boundaries. The solution space from continuous domain was reduced. Graph theory was utilized to construct a global barrier graph to model the pending problem, whether the directional sensors' sensing orientations are existed that can form strong barrier. A low-cost solution with the minimal number of directional sensors was designed as well. Simulation testifies the effectiveness of the proposed algorithm.
Traditional structured peer-to-peer networks especially distributed Hashing table (DHT) only support keyword-based exact search instead of semantic search. Although semantic DHT algorithms can implement semantic search by multi-routing-layers, they usually result in low precision ratio. A semantic similarity Hashing (SSH) scheme, based on standard alpha-stable distribution, was proposed to assign the similar identifiers to the resources with semantic similarity on Chord overlay. Consequently, the resources with semantic similarity could be stored in adjacent nodes of Chord overlay, so as to realize semantic search. Simulation shows that 70% recall ratio and 80% precision ratio are achieved among only 1.07% local nodes on an one-layer topology.
The performance of subspace algorithm which is based on eigenvalue decomposition degrades obviously, when gain and phase errors exist among component antennas of a dipole triad. To calibrate the gain and phase errors, a reference signal with unknown direction of arrival (DOA) is utilized in this paper. Two sets of data collected by the dipole triad before and after rotation around its x-axis are used. The closed-form solution of DOA and gain and phase errors are got by making use of the electromagnetic field ratio and DOA before and after rotation. Finally, the simulation results verify the efficacy of the proposed algorithm.
To improve the spatial sharing gain in millimeter-wave wireless personal network (WPAN), an interference-aware beamforming mechanism (IBF) was firstly proposed, and then a maximum coexistence possibility algorithm (MCP) based on physical interference model was presented. With help of interference information obtained using IBF and link coexistence conditions derived under physical interference model, MCP could increase the spatial sharing gain by efficiently and exactly identifying the coexistence link set. Simulations show that the MCP algorithm outperforms the maximum sector number difference (MSND) algorithm based on protocol interference model and the blind algorithm based on physical interference model in terms of spatial sharing gain, and the comparative advantage increases along with the sector number and node number.
The Doppler character of low-earth orbit satellites, carrying multi-beam receiving antenna, is different from the character of single-beam antenna systems. In order to analyze the Doppler character of each beam, expression of iso-Doppler characteristic curve was derived with help of establishment of vector analysis model. The characteristic curve picture was obtained using computer simulation. Analysis of each receiving multi-beam Doppler character was achieved when combined with the multi-beam coverage calculation. Finally, the preliminary application on the on-board multi-beam captures module was explained and illustrated.
Aiming at the conflict of the storage cost, the update frequency in P1 and the de-synchronizing defect in P2, an improved scalable radio frequency identification tag (RFID) ownership transfer protocol-lightweight scalable desynchronizing attack resistant protocol (LSDARP) was proposed. By the mechanism of sliding windows, the scheme only stores the tag's old and new adjacent eigenvalues to reduce the storage cost and avoid search efficiency degeneration when a group of eigenvalues was used out. It can be also re-synchronized when suffering de-synchronizing attacks continuously. Comparative analysis shows that LSDARP performs very well in aspects of the storage cost, computational complexity, backward security, de-synchronizing attack resistant and Tag Killing resistant.
A path balancing based ant colony optimization approach was presented. At first, in view of coordinated trajectory planning in dynamic and uncertain environments, the congestion factor was adopted to avoid collision between unmanned aerial vehicle (UAVs). Then, for time cooperation, the variable cooperative voyage was used to synchronize the arrival time of each UAV. With these premise of space-time collaborative, the trajectory for UAV fleet with minimum costs was selected. Simulation demonstrates the feasibility and the effectiveness of the proposed approach under complicated environment.
A multiple-input multiple-output fixed-complexity sphere decoder (FSD) with low-complexity channel matrix ordering is proposed based on reliabilities of the minimum mean square error (MMSE) decoder's decision variables. The proposed ordering algorithm first performs MMSE decoding, computes the reliabilities of MMSE decoder's decision variables, and finally orders channel matrix based on the reliabilities. The algorithm not only effectively reduces the complexity of FSD in preprocessing stage, but also overcomes its sequential nature, more suitable for applications. Simulation result shows that the performance loss caused by the low-complexity ordering algorithm is negligible and quasi-maximum likelihood performance can be achieved.
In order to improve the system capacity and simplify the receiver structure of traditional multi-carrier scenarios, a new orthogonal code called orthogonal time-frequency code was produced. Thereafter, a multi-carrier code division multiple access scheme was proposed. This new scheme employs orthogonal time-frequency code as its address code. It can not only improve the utilization of the codeword, but also can enhance the capacity of system. The frequency diversity gain in frequency selective fading channel can be obtained without coding by employed the Rake receive, that can also complete channel estimation and carrier recovery simultaneously. Simulation is coincided with the analysis.
For data collection in the presence of mobile sink and rendezvous points, a heuristic scheme is proposed jointly considering rendezvous selection, routing from sensors to rendezvous points and mobile sink path scheduling. The graph theory is used to model the network, and the minimum connected dominating set is selected as the initial rendezvous set. The final rendezvous set, the shortest path of mobile sink and the shortest route between sensors and rendezvous points are acquired through iteration procedure. The algorithm can reduce the data transmission from sensor nodes to rendezvous with guarantee of latency, thereby saving energy.
In order to maximize the spectrum utilization, the optimal transmission time of secondary user is derived according to primary user and secondary user's data transmission characteristics. An auto-adaptive spectrum sensing time allocation scheme based on partially observed Markov decision process (POMDP) was proposed. Execution of the best action based on the energy detection results and the belief state in the last decision slot was decided. Simulation claims that the proposed scheme can maximize spectrum utilization according to different collision probability constrain between primary users and secondary users.
To fulfill web service's requirements on trust evaluation and privacy protection, a privacy-aware trust model is put forward and a corresponding service selection mechanism is presented. They can help users to select the service providers which can provide better privacy protection besides service quality. In addition, to increase the generality of the proposed model, a kind of method based on correlation for attribute selection and weight measurement is designed which can adapt to different application environments. Finally, through experiments on data from real environment, the effectiveness of the proposed model is validated.
The energy efficiency (EE) optimization in uplink multi-user single-input multiple-output systems was addressed. The EE optimization model with minimum data rate requirement was first proposed, and the feasibility of model was analyzed. Then a convex approximation and geometric programming based energy-efficient joint beamforming and power control algorithm was developed. The convergence of the algorithm was proved. Simulation shows that the proposed algorithm can significantly improve EE performance and meet data rate requirements.
Ethernet ring protection (ERP) can provide ring automatic protection switch. A filtering database (FDB) flush was proposed based on G. 8032 ERP. Analysis and simulation show that the suggested scheme can effectively improve the performance of ethernet ring by decreasing unnecessary FDB flush in comparison with existing approaches.
In order to withstand and recover from concurrent multi-link failures in optical networks, a k-regular and k-connected structure based protection mechanism was proposed. Analysis shows that the k-regular and k-connected structure is an optimal protection structure in link-restorable mesh networks under multi-link failures. The linear programming model (ILP) for k-regular and k-connected structure and a construction method for k-regular and k-connected structure design were proposed. Simulations indicate that the spare capacity efficiency of k-regular and k-connected structure can reach the lower bound in static networks, and it can largely improve spare capacity efficiency compared with p-cycles based protection structure without reducing recovery rate under dynamic traffic.
Aiming at automatic extraction of relationships between domain concepts in ontology construction, a hybrid automatic extraction method is proposed. The relationships existing in domain concepts are classified into two kinds: one is the relationships with unknown types extracted by the extended association rule; the other is the relationships with the known types extracted by extraction rules. Meanwhile, the formation and expansion methods of extraction rules are presented. Experiments demonstrate that the proposed method is feasible and effective. It can not only obtain rich relationships in special domain, but also obtain better performance.
A dynamic trust model in mobile Ad hoc network (MANET) based on multiple factors decision (MDF-Trust) was proposed. Through quantifying the decision factors of state and behavior of the node, the routing node trust by multiple factors decision theory was evaluated. The deceptive behavior of malicious node was restrained by the trust model when introducing recommendation credibility and recommendation node searching algorithm. Simulation shows that the proposed trust model can reflect the node's trust changes accurately with better dynamic response ability with higher number of packet delivery and lower delay in the circumstance of existing deception nodes.
Dynamic wavelength and bandwidth assignment algorithm is a key issue for hybrid wavelength division multiplexing and time division multiplexing Ethernet passive optical network (WDM/TDM EPON). The influence of the heterogeneous round trip time of different optical network units (ONUs) was addressed and the problem was mapped as a multiprocessor scheduling problem with release dates. Preemptive shortest propagation delay (SPD)/ longest remaining processing time (LRPT) based scheduling algorithms named HSAⅠ和HSAⅡ in offline framework with the objective of minimizing the cycle length for the given tasks were proposed. Evaluation shows that HSAⅠ achieves shorter cycle length compared with the formal LRPT, SPD and lightweightoptimal wavelength scheduling (algorithm) (LOWSⅠ) algorithms, that is, achieves higher channel utilizations. For long-range PON, HSAⅡ and HSAⅠ achieve much better performances compared with the formal algorithms mentioned above.
In order to protect private information of users in the identity-based encryption (IBE) scheme, an anonymous IBE scheme in groups of prime order is proposed using asymmetric bilinear maps. Data security and receiver anonymity are realized. The attacker cannot obtain the receiver's identity from the ciphertext. Based on decisional bilinear Diffie-Hellman (DBDH) assumption, it is adaptive-ID secure without random oracles. The scheme in groups of prime order greatly reduces computational overhead compared with the ones in groups of composite order. The proposed scheme improves security and efficiency of anonymous IBE schemes simultaneously.
The present direction of arrival (DOA) estimation methods in impulsive noise are mostly based on the fractional lower order statistics, and the computational load is much heavier than the traditional methods based on the second-order moment. Utilizing the property of the impulsive noise, a new DOA estimation method is proposed. Firstly, the reference amplitude of the array receiving data is estimated, which is not influenced by the impulsive noise. Then, the data with amplitude higher than the reference amplitude is normalized. Finally, the common second-order moment method is exploited to estimate the DOAs. Theoretical analysis and simulation results show that the proposed method has small computational load, and is effective in the presence of strong impulsive noise.
In order to improve the accuracy and controllability of resource allocation process in multiple virtual planes, a prediction-based adaptive allocation algorithm is proposed. The algorithm used a grey prediction model to predict the resource usage for each virtual node at the next time, and then adaptively distributed and modified the value of resource allocation through the establishment of the mathematical model between the allocation factors and resource utilization, until it could meet the actual demand. For different virtual business plane, the algorithm could set different allocation parameters to control the allocation results so as to meet the different levels of service. Finally, the experiments show that the algorithm can not only implement the process of dynamic resource allocation quickly and accurately, but also meet the distribution of multiple virtual business planes. And in this process, the virtual resource utilization is maintained at a reasonable value near the target.
A clock edge controllable dual edge-triggered D flip-flop is presented using the cellular neural network (CNN). Based on analysis of the relevance of uncoupled CNN and linear separable boolean function, a fundamental approach to implement logical circuit with uncoupled CNN is described. Implementation of the new D flip-flop circuit also is discussed. Different from traditional D flip-flop, this new method is a reconfigurable component, which can implement the up edge-triggered D flip-flop; the down edge-triggered and the dual edge-triggered respectively only by changing the gene parameters of CNN. Thus, a new method to design the reconfigurable logical circuit is created.