In this study, a hierarchical architecture for the intermittent control under the minimum transition hypothesis(MTH) was implemented. A two-stage hierarchy was adopted to perform the high-level and the low-level con...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
In this study, a hierarchical architecture for the intermittent control under the minimum transition hypothesis(MTH) was implemented. A two-stage hierarchy was adopted to perform the high-level and the low-level control respectively. The high-level controller performed the intermittent control by setting a sequence of goals for the low-level controller. Goal planning as the intermittent control policy was learned with hierarchical deep deterministic policy gradient(h-DDPG) proposed in this study, which is a hierarchical version of the conventional DDPG. The model successfully learned to temporally decompose a complex movement into a sequence of basic motor skills with sparse transitions, as shown in results of the two validation experiments: the trajectory following and the obstacle avoidance tasks.
Recently,great progesses have been made in using discriminative classifiers in object *** specifically,correlation filters(CFs) for visual tracking have been attractive due to t heir competitive performances on both a...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
Recently,great progesses have been made in using discriminative classifiers in object *** specifically,correlation filters(CFs) for visual tracking have been attractive due to t heir competitive performances on both accuracy and *** this paper,the latest and representative approaches of CF b ased trackers are presented in d *** addition,trackers used deep convolutional features are introduced and several famous tracking methods which fine-tune the pretrained deep network are *** evaluate the performances of different trackers,a detailed introduction of the evaluation methodology and the datasets is described,and all introduced trackers are compared based on the mentioned ***,several promising directions as the conclusions are drawn in this paper.
The power industry innovation has increasingly become a top concern for current reforms. Power systems feature scattered data storage, incapable data analysis ability, poor computing capability, and ineffective intera...
详细信息
ISBN:
(纸本)9781509034840
The power industry innovation has increasingly become a top concern for current reforms. Power systems feature scattered data storage, incapable data analysis ability, poor computing capability, and ineffective interaction interface. To resolve these issues, we need multiple data mining techniques to extract information for analytical capacity improvement. Secondly, we need visualization techniques to analyze and optimize interaction. Lastly, we need distributed technologies for unified data management to increase computing capability and system scalability. Considering China's smart grid information, this paper proposes solutions to problems, such as the existing underdeveloped power management systems, a lack of automation methods, low data visualization, and poor data management. The electric power industry has functional requirements for this research. Based on existing data mining, visualization and understanding of distributed technologies, we discussed the functions of each part of the implementation in a smart grid management system: the data mining module, visualization module and data management module.
This paper analyzes the complexity of Internet innovation using scale-free network theory. The evolutionary model has also been numerically simulated. The conclusion is that the model has small-world and scale-free fe...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
This paper analyzes the complexity of Internet innovation using scale-free network theory. The evolutionary model has also been numerically simulated. The conclusion is that the model has small-world and scale-free features. The small-world effect shows that there is a wide range of high efficiency Internet resource integration. Scale-free features indicate that a few core nodes become central. Therefore, we should not only see the characteristics to enhance the positive effect of innovation performance, but also avoid the risk of lockin and vulnerability.
Offloading computation-intensive tasks from mobile to nearby resource-rich surrogates, called edge servers, is proposed recently because traditional mobile cloud computing has a bottleneck of bandwidth and resource li...
详细信息
ISBN:
(纸本)9781538676721
Offloading computation-intensive tasks from mobile to nearby resource-rich surrogates, called edge servers, is proposed recently because traditional mobile cloud computing has a bottleneck of bandwidth and resource limitation for devices. The primary performance concern of offloading is how to maximize energy saving under task delay and task dependency limitation. Besides, edge servers that mobile perceived are changeable and heterogeneous in the process of offloading. In this paper, we formalize this problem, reduce it into knapsack problem and propose a task scheduling scheme, named TaSRD, including independent sub-task scheduling for tasks without dependencies and dependent sub-task scheduling for dependent tasks. We implement TaSRD and evaluate it by case study and simulation on CloudSim framework developed by Melbourne University. We use time model and energy model to measure results and recommend suitable parameters for TaSRD. The experimental results demonstrate that TaSRD can effectively save energy and reduce makespan for mobile while offloading tasks to edge servers.
The task assignment problem for multiple vessels cooperative driving is the key problem of the multiple vessels cooperative control. Due to the system showing multi-objective, multi-tasking and multi-constrained featu...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
The task assignment problem for multiple vessels cooperative driving is the key problem of the multiple vessels cooperative control. Due to the system showing multi-objective, multi-tasking and multi-constrained features, a cooperative multi-task assignment model is proposed, which can transform multiple constraints task assignment problem into multiple constraints optimization problem based on the multiple vessels task assignment cost function. This method optimizes the results of task assignment by using genetic ant colony hybrid algorithm to search optimization solution globally. In the simulation experiment, it is compared with the genetic algorithm and the ant colony algorithm alone, and experimental results show that the method can optimize task assignment on the basis of satisfying these constraints.
Hyperspectral image(HSI) is often contaminated by mixed noise in the acquisition process. In this paper,a hyperspectral image low-rank restoration method based spectral-spatial total variation(LRSSTV) is proposed....
详细信息
ISBN:
(纸本)9781538619797;9781538619780
Hyperspectral image(HSI) is often contaminated by mixed noise in the acquisition process. In this paper,a hyperspectral image low-rank restoration method based spectral-spatial total variation(LRSSTV) is proposed. The spectral high correlation is exploited by low-rank representation and the sparse noise is represented by the l-norm. Furthermore,to remove the Gaussian noise and enhance the edge information,spectral-spatial adaptive total variation prior knowledge is utilized. Both simulated and real-world data experimental results show that the proposed method can work well in detail preservation and noise removal.
Recently, brain-computer interface has been applied to many fields such as steady-state visual evoked potential(SSVEP). However, in the conventional method, the frequency resolution is low due to the dependence of the...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
Recently, brain-computer interface has been applied to many fields such as steady-state visual evoked potential(SSVEP). However, in the conventional method, the frequency resolution is low due to the dependence of the short-time Fourier transform on the analysis window length. Therefore, it is not possible to analyze a non-integer multiple signal, as a side-lobe will occur. We verified the precision of non-harmonics analysis,and proposed and attempted to analyze the change and stimulus of SSVEP. We found the frequency resolution to be improved exponentially.
This paper proposed a tracking algorithm based on sparse coding and spectral residual saliency under the framework of particle filtering. The proposed algorithm can be divided into three parts. Firstly, spectral resid...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
This paper proposed a tracking algorithm based on sparse coding and spectral residual saliency under the framework of particle filtering. The proposed algorithm can be divided into three parts. Firstly, spectral residual is used to calculate a saliency map of the current frame and then compute the saliency score of each particle. Secondly, several particles are eliminated directly based on the differences between the saliency scores of the particles in the current frame and the target score in the prior frame. Thirdly, Sc SPM is used to compute the observation vector for the rest particles and the tracking task is finished in the framework of particle filtering. Both quantitative and qualitative experimental results demonstrate that the proposed algorithm performs favorably against the nine state-ofthe-art trackers on ten challenging test sequences.
In this paper, we propose an online visual tracking algorithm for fused sequences via deep learning and adaptive Particle filter(PF). Our algorithm pretrains a simplified Convolution Neural Network(CNN) to obtain a ge...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
In this paper, we propose an online visual tracking algorithm for fused sequences via deep learning and adaptive Particle filter(PF). Our algorithm pretrains a simplified Convolution Neural Network(CNN) to obtain a generic target representation. The outputs from the hidden layers of the network help to form the tracking model for an online PF. During tracking, the moving information guides the distribution of particle samples. The tests illustrate competitive performance compared to the state of-art tracking algorithms especially when the target or camera moves quickly.
暂无评论