Journal of Information Science and Engineering
- OpenAccess
社團法人中華民國計算語言學學會,正常發行
選擇卷期
- 期刊
- OpenAccess
An accurate frequency synthesizer is essential in wireless communications, radar systems, and frequency metrology. However, open-loop signal sources exhibit severe frequency fluctuation and are vulnerable to supply-induced frequency drift, phase noise, power consumption. There is a demand for precise oscillation frequency with wide tuning range and low phase noise. This motivates the proposed synthesizer to achieve relatively lower in-band phase noise as well as good out-of-band phase noise through the use of digital amplitude control circuit. This paper presents a low power, low phase noise, and fast locking CMOS PLL frequency synthesizer. The frequency synthesizer is designed by using the 65nmCMOS technology. It can support LTE, GSM/EDGE application with the frequency ranged from 4.39 GHz to 5.71 GHz for the local oscillator in the RF front-end circuits. This paper achieves the faster locking with the lock check through controlling the phase detector and charge pump to enhance the locking speed of the proposed PLL. By implementing the proposed design, the locking speed can be enhanced along with minimum power consumption and phase noise.
- 期刊
- OpenAccess
Device-free indoor human trajectory tracking is critical to support health care applications for elderly people. Many device-free localization algorithms depend on expensive hardware to achieve tracking accuracy. In contrast to such algorithms, this paper proposes a new device-free human trajectory tracking algorithm for indoor environments based on channel state information that is extracted from a Wi-Fi network interface card, which is a low-cost component. The proposed algorithm first uses the characteristics of locally linear embedding to detect whether a person is moving and applies quadratic discriminant analysis to determine the new location of the person. The determined locations of the person are connected to form a trajectory. Experimental results revealed that the proposed algorithm provides an effective solution for passive human trajectory tracking.
- 期刊
- OpenAccess
Mobile station (MS) localization often suffers from hybrid line of sight (LOS), one-bound (OB) and multiple-bound (MB) non-line of sight (NLOS) propagation in multipath environments. Due to the unknown propagation path, accurate position estimate of MS is challenging through using the measured angle of departure (AOD), angle of arrival (AOA), and time of arrival (TOA) of signal between MS and base station (BS). To address this problem, a new weighting localization algorithm based on LOS and OB NLOS identification is proposed in this paper. For each propagation path, by utilizing the geometric relation between AOD and AOA, a theoretic threshold is derived to decide whether it is LOS or NLOS propagation. Moreover, in order to further discriminate OB or MB NLOS propagation, an effective cost function is formulated and an iterative OB NLOS identification method is proposed to discard MB NLOS propagation paths. Finally, a weighting localization algorithm is applied for fusing the measured data of LOS and OB NLOS propagation paths. Simulation results demonstrate that simulation of LOS identification method is consistent with theoretic one, and the proposed algorithm can greatly improve the localization accuracy of MS in different multipath environments, especially when LOS path is available.
- 期刊
- OpenAccess
This paper investigates the impact of node mobility and imperfect channel state information (CSI) on the end-to-end performance of a selective decode-and-forward (S-DF) based multiple-input multiple-output (MIMO) space-time block-code (STBC) cooperative wireless system. A closed form expression is derived for the per-block average pair-wise error probability (PEP) for several configurations in terms of number of phases, hops and relays over time selective Rayleigh fading channel, with best relay selection (BRS) and path selection (PS). Further, a framework is developed for deriving the diversity order (DO) for each configuration. Results show that when both destination node (DN) and source node (SN) are immobile, system performance does not encounter asymptotic error floor although relay node (RN) is mobile. Although with mobile RN, the movement of either the DN or the SN critically affects system performance by asymptotic error floors. System performance is analyzed for both equal power and optimal power scenario and the results show that system performance improves with optimal power. Simulation results are in close agreement with the analytical results at high signal to noise ratio (SNR) regimes.
- 期刊
- OpenAccess
NMF has been extensively applied on various pattern recognition problems, including face recognition. To enhance the performance of NMF for face recognition, new additive iterative step sizes are proposed for the basic NMF method, which can raise the searching accuracy during the iteration, and then the recognition rate can be improved. The improved NMF method is named INMF. Meanwhile, the experiments results show that the proposed improved additive iteration can also raise the recognition rate of the SNMF and WNMF. Besides, we find that no sparse constraint is applied to INMF and lots of redundant information still exists, thus a threshold-sparse constraint is introduced to make the base matrix W to a 0-1 matrix, and then the feature data of the base matrix become sparse, therefore the recognition rate can be further improved. The INMF model with the threshold-sparse constraint is named SINMF. Finally, our extensive experimental results showed that the highest recognition rate of the SINMF method can achieve 99%, with improvement over the INMF, IWNMF, ISNMF and deep NMF methods by 11%, 5.5%, 11% and 8%, respectively. Meanwhile, compared with the deep convolutional neural network, the recognition rate of SINMF method is proved more efficient.
- 期刊
- OpenAccess
Compared with the traditional static traveling salesman problem, it is more practical to study the dynamic traveling salesman problem in dynamic environment. In this paper, the dynamic optimization problem is considered as a combination of a series of static optimization problems, and an adaptive ant colony algorithm is proposed to solve the dynamic traveling salesman problem. When traveling salesman problem changes, we firstly analyze the environmental changing degree of traveling salesman problem, then an adaptive pheromone initialization mechanism adaptable to changing degree is presented, so as to ensure faster convergence speed without affecting the accuracy of problem solving. In addition, to further improve the quality of solution, we propose an optimization guidance based search mechanism and integrate it into the adaptive ant colony algorithm proposed. Finally, the algorithm is analyzed and compared with the related algorithms on several common real data sets. The results show that the adaptive ant colony algorithm proposed in this paper can solve the dynamic traveling salesman problem more effectively.
- 期刊
- OpenAccess
Spatio-temporal data modeling is an important basis for spatio-temporal data management. Unified Modeling Language (UML) is a widely used modeling language. Therefore, how to model spatio-temporal data based on UML and then how to further verify the correctness of the spatio-temporal UML models have become important issues. In this paper we propose a spatio-temporal UML model and a Description Logic (DL) method for verifying the model. First of all, we present a UML-based spatio-temporal data model. Also, an abstract definition and semantic description of the spatio-temporal UML models are given, and a case of cadastral change process is provided. Then, by adding some special concepts, roles, and axioms into the DL ALCIQ, a method for mapping the spatio-temporal UML models to ALCIQ knowledge bases is proposed, and a mapping example is provided. Further, several verification tasks of the spatio-temporal UML models are equivalently converted to the inference problems of the mapped ALCIQ knowledge bases, and the inference results can be returned and the verification of spatio- temporal UML models are realized with the help of the DL inference abilities.
- 期刊
- OpenAccess
In recent years, due to the great potential of genetic algorithms to solve complex optimization problems, it has attracted wide attention. But the traditional genetic algorithm still has some shortcomings. In this paper, a new adaptive genetic algorithm (NAGA) is proposed to overcome the disadvantages of the traditional genetic algorithm (GA). GA algorithm is easy to fall into the local optimal solution and converges slowly in the process of function optimization. NAGA algorithm takes into accounts the diversity of the population fitness, the crossover probability and mutation probability of the nonlinear adaptive genetic algorithm. In order to speed up the optimization efficiency, the introduced selection operator is combined with the optimal and worst preserving strategies in the selection operator. And in order to keep the population size constant during the genetic operation, the strategy of preserving the parents is proposed. Compared with the classical genetic algorithm GA and IAGA, the improved genetic algorithm is easier to get rid of the extremum and find a better solution in solving the multi-peak function problem, and the convergence rate is faster. Therefore, the improved genetic algorithm is beneficial for function optimization and other optimization problems.
- 期刊
- OpenAccess
In this paper, under the condition of D-differentiation, we consider the fuzzy programming problem with the general fuzzy mapping (non-convex) as the objective mapping. By discussing the characteristics of the optimal solution of unconstrained fuzzy programming, we give the KKT condition of the optimal solution of more general fuzzy programming with real value function as the constrained condition, and some test examples. Meanwhile, we discuss the optimal condition of a special class of fuzzy programming problem with the real-valued concave function as the constrained condition and the convex fuzzy mapping as the objective mapping.
- 期刊
- OpenAccess
This paper focuses on the state estimation problem of target tracking with intermittent measurements. Leveraged by the posterior measurements, an amended Kalman filter is proposed in this paper to improve the precision of the current estimated state. Both the deduction and proof of the amended Kalman filter are discussed specifically to distinguish amended Kalman filter from the Kalman smoother. Extensive simulations are conducted and the simulation results verify the excellent tracking performance of the amended Kalman filter.