Artificial Intelligence Key Laboratory of Sichuan Province

Zigong, China

Artificial Intelligence Key Laboratory of Sichuan Province

Zigong, China
Time filter
Source Type

Li T.,University of Sichuan | Wang Y.,University of Sichuan | Wang Y.,Artificial Intelligence Key Laboratory of Sichuan Province | Yang Y.,University of Sichuan
Discontinuity, Nonlinearity, and Complexity | Year: 2016

In this letter stability analysis of fractional order nonlinear systems is studied. An extension of Lyapunov direct method for fractional order systems is proposed by using the properties of Mittag-Leffler function and Laplace transform. Some new sufficient conditions which ensure local exponential stability of fractional order nonlinear systems are proposed firstly. And we apply these conditions to the Riemann-Liouville fractional order systems by using fractional comparison principle. Finally, three examples are provided to illustrate the validity of the proposed approach. © 2016 L&H Scientific Publishing, LLC. All rights reserved.

Cao L.,University of Sichuan | Cao L.,Artificial Intelligence Key Laboratory of Sichuan Province | Hu X.,Xi'an Research Institute of High Technology | Zhang S.,Xi'an Research Institute of High Technology | And 2 more authors.
Journal of Aerospace Engineering | Year: 2017

The sensor-based backstepping control law, which is not sensitive to model uncertainty, uses less model knowledge and the measurements of state derivatives. A flight control law based on a sensor-based backstepping method for unmanned aerial vehicles (UAVs) is proposed in this paper. The nonlinear model uncertainties of a UAV and sensor noise are concerning. Moreover, command filters are used to handle dynamics and saturation effects of the actuators for carrying out the flight control law. The simulation results show the proposed flight control law can guarantee the closed-loop system stability and has better tracking control performance than the block backstepping control law with model uncertainty. © 2017 American Society of Civil Engineers.

Wu D.-Q.,Donghua University | Wu D.-Q.,University of South China | Wu D.-Q.,Artificial Intelligence Key Laboratory of Sichuan Province | Zheng J.-G.,Donghua University
Kongzhi yu Juece/Control and Decision | Year: 2013

In view of the limitation of the current kinds of particle swarm optimization(PSO) algorithm, a self-adaptive learning of hybrid strategy algorithm based on parallel particle swarm optimization(HLPSO) is proposed. The algorithm combines four strategies reasonably in the different point of view: Convergence, jump out, exploration and exploitation, which chooses an appropriate strategies to solve the different forms of problems through adjusting the probability of the strategies gradually in the process of optimizing. Moreover, simulation experiment on a suite of 7 benchmark functions is given, and the comparisons with other algorithms are provided. The results show that the proposed approach has better convergence rate and great capability of preventing premature convergence.

Deng W.,Dalian Maritime University | Deng W.,Dalian Jiaotong University | Deng W.,Chongqing University of Technology | Deng W.,Artificial Intelligence Key Laboratory of Sichuan Province | And 5 more authors.
Computers and Mathematics with Applications | Year: 2012

A novel parallel hybrid intelligence optimization algorithm (PHIOA) is proposed based on combining the merits of particle swarm optimization with genetic algorithms. The PHIOA uses the ideas of selection, crossover and mutation from genetic algorithms (GAs) and the update velocity and situation of particle swarm optimization (PSO) under the independence of PSO and GAs. The proposed algorithm divides the individuals into two equation groups according to their fitness values. The subgroup of the top fitness values is evolved by GAs and the other subgroup is evolved by the PSO algorithm. The optimal number is selected as a global optimum at every circulation which shows better results than both PSO and GAs, then improves the overall performance of the algorithm. The PHIOA is used to optimize the structure and parameters of the fuzzy neural network. Finally, the experimental results have demonstrated the superiority of the proposed PHIOA to search the global optimal solution. The PHIOA can improve the error accuracy while speeding up the convergence process, and effectively avoid the premature convergence to compare with the existing methods. © 2011 Published by Elsevier Ltd. All rights reserved.

Zhang Z.,Jiangsu University of Science and Technology | Yang X.,Jiangsu University of Science and Technology | Yang X.,Nanjing University of Science and Technology | Yang X.,Artificial Intelligence Key Laboratory of Sichuan Province
Frontiers of Computer Science | Year: 2014

Presently, the notion of multigranulation has been brought to our attention. In this paper, the multigranulation technique is introduced into incomplete information systems. Both tolerance relations and maximal consistent blocks are used to construct multigranulation rough sets. Not only are the basic properties about these models studied, but also the relationships between different multigranulation rough sets are explored. It is shown that by using maximal consistent blocks, the greater lower approximation and the same upper approximation as from tolerance relations can be obtained. Such a result is consistent with that of a single-granulation framework. © 2014, Higher Education Press and Springer-Verlag Berlin Heidelberg.

Gao S.,Jiangsu University of Science and Technology | Yu H.,Jiangsu University of Science and Technology | Qiu L.,Artificial Intelligence Key Laboratory of Sichuan Province | Cao C.,CAS Institute of Computing Technology
International Journal of Computers and Applications | Year: 2014

According to the idea of wading across the stream by feeling the way, a kind of fast efficient random optimization algorithm is proposed. The wading across stream algorithm (WSA) acts as a solution as a start point, then searches several random solutions near the start point, and finds the best of these solutions. This best solution is taken as the next start point, and then several random solutions near this start point are searched, and so on. For solving continuous optimization problem, the improved wading across stream algorithm (IWSA) gradually shrinks the search space. The experimental results of some classic benchmark functions show that the proposed optimization algorithms improve extraordinarily the convergence velocity and precision. For solving the travelling salesman problem, the improved method selected the best of the initial solution as the start solution. To search the neighbourhood trial solution, four strategies are proposed. It is proved that reversal strategy is a simple and effective algorithm.

Song Y.,Dalian Maritime University | Chen R.,Dalian Maritime University | Liu Y.,Dalian Maritime University | Liu Y.,Artificial Intelligence Key Laboratory of Sichuan Province
Journal of Computers (Finland) | Year: 2012

The Semantic Web is the extension of the World Wide Web that enables people to share content beyond the boundaries of applications and websites. The understanding of Semantic Web documents is built upon ontologies that define concepts and relationships of data. Hence, the correctness of ontologies is vital. In this paper, we propose a new algorithm combined with the software engineering techniques, such as Alloy modeling language and its reasoner Alloy Analyzer to provide checking and reasoning service for OWL ontologies. First of all, we use Jena to parse OWL ontology documents. Next, the intermediate results are used as the inputs of the algorithms to generate the Alloy model. Futher, with the assistance of Alloy Analyzer, the Alloy model is checked. Experimental results show that this method can be carried out large-scale ontology reasoning and complex-property reasoning which are different from traditional ontology reasoning. Furthermore, the results provide useful information to guide the ontology modification. © 2012 ACADEMY PUBLISHER.

Li K.,University of Sichuan | Zeng H.,Artificial Intelligence Key Laboratory of Sichuan Province
Mathematics and Computers in Simulation | Year: 2010

In this paper, we investigate a class of impulsive Cohen-Grossberg-type BAM neural networks with time-varying delays. By establishing the delay differential inequality with impulsive initial conditions, and employing the homeomorphism theory, the M-matrix theory and the inequality a∏k=1lbkqk≤(1r)(ar+∑k=1lqkbkr) (a≥0,bk≥0,qk≥0 with ∑k=1lqk=r-1, and r≥1), some new sufficient conditions ensuring the existence, uniqueness and global exponential stability of equilibrium point for impulsive Cohen-Grossberg-type BAM neural networks with time-varying delays are derived. In particular, the estimate of the exponential convergence rate which depends on the system parameters and the impulsive disturbance intension is also provided. An example is given to show the effectiveness of the results obtained here. © 2010 IMACS.

Qiu L.,University of Sichuan | Wei Z.X.,University of Sichuan | Liu Y.,Artificial Intelligence Key Laboratory of Sichuan Province
Applied Mechanics and Materials | Year: 2014

RFID middleware is a key technology of Internet of things. It achieves data recognition and data filtering between hardware devices and software applications. A design scheme for RFID middleware framework is presented in the paper. According to functional analysis and layered design ideas, we define a hierarchy of RFID middleware and construct an application framework, in which RFID event manager module is responsible for processing data streams and RFID information service module is responsible for system integration. Simulation results verify the correctness of the design project. Its recognition rate, redundant data filtering capability and other performance indicators have reached the preset requirements. © (2014) Trans Tech Publications, Switzerland.

Huang D.,University of Sichuan | Huang D.,Artificial Intelligence Key Laboratory of Sichuan Province | Yu S.,University of Sichuan | Tian J.,University of Sichuan | Hu Y.,University of Sichuan
Yi Qi Yi Biao Xue Bao/Chinese Journal of Scientific Instrument | Year: 2016

Aiming at the deficiencies of high cost, inconvenience in powering the positioning system and poor scalability of the existing indoor positioning systems, this paper proposes an indoor positioning system based on power-line; and the working principle of the system is discussed. On the basis of above the hybrid indoor positioning algorithm based on Support Vector Classification Machine and K-Nearest Neighbor is proposed. The principle of the algorithm is discussed in detail. Through the signal collection with the acquisition system, the indoor transmission characteristics of the positioning signal based on power-line is analyzed and the RSSI feature sample library is established. At last, the positioning experiments were conducted using SVCM, KNN and SVCM-KNN algorithms; their positioning performances were compared. The experiment results show that the SVCM-KNN algorithm can reduce the position error effectively and achieve the indoor positioning precision requirement. © 2016, Science Press. All right reserved.

Loading Artificial Intelligence Key Laboratory of Sichuan Province collaborators
Loading Artificial Intelligence Key Laboratory of Sichuan Province collaborators