This paper proposes a reinforcement learning based tag recommendation algorithm to deal with the data sparseness that affects the performance stability of collaborative filtering algorithms. Our algorithm integrates u...
详细信息
作者：
Haibo DuShihua LiSchool of Automation
Southeast University Nanjing 210096 P.R.China Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education
Abstract This paper investigates the finite-time attitude stabilization problem for a rigid spacecraft with an unknown inertia matrix. By using homogeneous approach, a local continuous finite-time controller is first ...
Abstract This paper investigates the finite-time attitude stabilization problem for a rigid spacecraft with an unknown inertia matrix. By using homogeneous approach, a local continuous finite-time controller is first designed such that the attitude can be stabilized to the equilibrium in finite time. Then, to solve global attitude stabilization problem, a modified finite-time controller is proposed. Rigorous proof shows that the attitude can be stabilized to the equilibrium in finite time. Finally, a simulation example is given to demonstrate the efficiency of the proposed method.
Abstract In this paper, the speed-regulation problem for permanent magnet synchronous motor (PMSM) system under vector control framework is studied. For the speed loop, a standard internal model controller is first de...
Abstract In this paper, the speed-regulation problem for permanent magnet synchronous motor (PMSM) system under vector control framework is studied. For the speed loop, a standard internal model controller is first designed based on a first-order model of PMSM by analyzing the relationship between reference quadrature axis current and speed. Considering the disadvantages that the standard IMC method is sensitive to control input saturation and provides a poor load disturbance rejection performance, a modified IMC is introduced where a feedback control term is added to form a composite control structure. For the two current-loops, PI algorithms are employed respectively. Simulation and experimental results show the effectiveness of the proposed control method.
To design an energy efficient data collection scheme instead of multi-hop data relay is essential for Wireless Sensor Networks. To many proposed data gathering approaches, the long data delay is the main problem. In t...
详细信息
To design an energy efficient data collection scheme instead of multi-hop data relay is essential for Wireless Sensor Networks. To many proposed data gathering approaches, the long data delay is the main problem. In this paper, mobile collectors are introduced in a sensor network and a data collection mechanism is proposed. To organize sensors into three-level clusters, an energy efficient clustering algorithm is proposed. The collection strategy is based on the hierarchical structure and consists of two rules to schedule mobile collectors reasonably. Simulation results demonstrate that the data collection scheme effectively reduces the energy consumption and decrease the data delay.
A recent experimental study [Traulsen et al., Proc. Natl. Acad. Sci. 107, 2962 (2010)] shows that human strategy updating involves both direct payoff comparison and the cost of switching strategy, which is equivalent...
详细信息
A recent experimental study [Traulsen et al., Proc. Natl. Acad. Sci. 107, 2962 (2010)] shows that human strategy updating involves both direct payoff comparison and the cost of switching strategy, which is equivalent to inertia. However, it remains largely unclear how such a predisposed inertia affects 2×2 games in a well-mixed population of finite size. To address this issue, the “inertia bonus” (strategy switching cost) is added to the learner payoff in the Fermi process. We find how inertia quantitatively shapes the stationary distribution and that stochastic stability under inertia exhibits three regimes, with each covering seven regions in the plane spanned by two inertia parameters. We also obtain the extended “1/3” rule with inertia and the speed criterion with inertia; these two findings hold for a population above two. We illustrate the above results in the framework of the Prisoner's Dilemma game. As inertia varies, two intriguing stationary distributions emerge: the probability of coexistence state is maximized, or those of two full states are simultaneously peaked. Our results may provide useful insights into how the inertia of changing status quo acts on the strategy evolution and, in particular, the evolution of cooperation.
Automatic service composition is a key issue of cooperative resource utilization in component-based networked robot system. In this paper, a semantic-based automatic service composition method is proposed. The semanti...
详细信息
Automatic service composition is a key issue of cooperative resource utilization in component-based networked robot system. In this paper, a semantic-based automatic service composition method is proposed. The semantic representation and composition of service is achieved by semantic descriptions of service functions and formulating a set of I/O parameters matching rules. Based on this, a bidirectional breadth-first traversal search algorithm inspired by the maze problem is proposed. As a result, the proposed method is capable of automatically generating plans to compose the functions according to the task requirement. Experimental results validate the effectiveness and the flexibility of the proposed method.
In terms of the difficulty of vehicle tracking in complex environment of the visual surveillance system, an object tracking algorithm is proposed for the applications in practical visual surveillance systems for intel...
详细信息
In terms of the difficulty of vehicle tracking in complex environment of the visual surveillance system, an object tracking algorithm is proposed for the applications in practical visual surveillance systems for intelligent traffic. A block-based Gaussian mixture background modeling method for object detection is presented to reduce the computational complexity of moving vehicle object abstraction. An adaptive tracking algorithm fused with color features and texture features is described to better adapt the traffic scene variation. The experimental results show that the proposed algorithm can effectively deal with the complex urban traffic conditions and the tracking performance is better than the conventional particle filter method and single feature based non-adaptive object tracking method.
A novel image deblurring method based on high-order non-local range Markov Random Field (NLR-MRF) prior is proposed in the paper. NLR-MRF provides an effective framework to model the statistical prior of natural image...
详细信息
A novel image deblurring method based on high-order non-local range Markov Random Field (NLR-MRF) prior is proposed in the paper. NLR-MRF provides an effective framework to model the statistical prior of natural images and leads to excellent performance in the application of image denoising and inpainting. Moreover, the framework will be extended to image deblurring in our work. Instead of commonly used maximum a-posteriori (MAP) estimation, which has several shortcomings, the high-order NLR-MRF prior is integrated into Bayesian minimum mean squared error (MMSE) estimation framework. Then, an efficient Gibbs sampling algorithm is adopted to compute MMSE estimation. The proposed method frees the user from determining regularization parameter beforehand, which relies on unknown noise level. We perform experiments on synthetic and real-world data to demonstrate the effectiveness of our method. Both quantitatively and qualitatively evaluations show superior or comparable results to the state-of-art deblurring methods.
Robust real-time tracking of non-rigid objects is a challenging task. Color is a powerful feature for tracking deformable objects in image sequences with complex backgrounds. Color distribution is applied, as it is ro...
详细信息
Robust real-time tracking of non-rigid objects is a challenging task. Color is a powerful feature for tracking deformable objects in image sequences with complex backgrounds. Color distribution is applied, as it is robust to partial occlusion, is rotation and scale invariant and computationally efficient. Particle filter has been proven very successful for non-linear and non-Gaussian estimation tracking problems. The article presents the integration of color distributions into particle filtering. A target is tracked with a particle filter by comparing its histogram with the histograms of the sample positions using the Bhattacharyya distance. Additionally, to solve the sample impoverishment (all particles collapse to a single point within a few iterations) in the particle-filter algorithm, a new resampling algorithm is proposed to tackle sample impoverishment. The performance of the proposed filter is evaluated qualitatively on various real-world video sequences. The experimental results show that the improved color-based particle filter algorithm can reduce sample impoverishment effectively and track the moving object very well.
Emergy analysis provides a feasible approach to evaluate the status and position of different energy carriers in the universal energy hierarchy. In this paper, an emergy-based method is conducted to measure the resour...
详细信息
Emergy analysis provides a feasible approach to evaluate the status and position of different energy carriers in the universal energy hierarchy. In this paper, an emergy-based method is conducted to measure the resources input of Chinese industry from 1997 to 2006. Resources inflows including fossil fuels, mineral resources, agricultural products, and other imported materials are accounted, based on which related indicators including resources intensity, industrial output, and environmental emissions are investigated. Results show a steady upward trend for the total resources input of Chinese industry during the past decade. The total resources input amounted to 1.53×1025 sej in 2006, of which non-renewable one accounted for 70.65% owing to the dominating input of fossil fuels and nonmetal minerals. Resources intensity measured by the ratio of resources input to industrial value added declined gradually during 1997–2002, but the rapid expansion of resource–intensive sub–sectors resulted in a reverse trend since 2003. The current resources use pattern of Chinese industry is characterized by increasing input of non-renewable resources, excessive expansion of resource–intensive production, and tremendous challenge from environmental pressure.
暂无评论