Skip to main content
main-content

Über dieses Buch

This book provides scientific research into Cognitive Internet of Things for Smart Society, with papers presented at the 2nd EAI International Conference on Robotic Sensor Networks. The conference explores the integration of networks and robotic technologies, which has become a topic of increasing interest for both researchers and developers from academic fields and industries worldwide. The authors posit that big networks will be the main approach to the next generation of robotic research, with the explosive number of networks models and increasing computational power of computers significantly extending the number of potential applications for robotic technologies while also bringing new challenges to the network's community. The 2nd EAI International Conference on Robotic Sensor Networks was held 25-26 August 2018 at the Kitakyushu International Conference Center (MICE), Kitakyushu, Japan.

Inhaltsverzeichnis

Frontmatter

New Tuning Formulas: Genetic Algorithm Used in Air Conditioning Process with PID Controller

Abstract
In this paper, a new tuning formula is proposed for PID controller in air conditioning processes, which is often used in engineering practice. A qualified controller should achieve a good balance between system performance and robustness. In this study, the set-point following and attenuation of load disturbances are decoupled by using two degrees-of-freedom control structure and the ability to reject load disturbances to the robustness constraints is maximized by optimization algorithms. A novel scheme is adopted in the derivation of the new tuning formulas that can be much simpler and easier with similar control performance compared to complex optimization algorithms. The simulation results are given to demonstrate their feasibility and effectiveness.
Xiaoli Qin, Hao Li, Weining An, Hang Wu, Weihua Su

A Multi-Level Thresholding Image Segmentation Based on an Improved Artificial Bee Colony Algorithm

Abstract
As a popular evolutionary algorithm, artificial bee colony (ABC) algorithm has been successfully applied into the threshold-based image segmentation problem. Based on our analysis, we find that the Otsu segmentation function is separable which means each variable is independent. Due to its one-dimensional search strategy and relative power global but poorer local search abilities, ABC could find an acceptable but not precise segmentation results. For making more precise search and further enhancing the achievements on image segmentation, we propose an Otsu segmentation method based on a new ABC algorithm with an improved scout bee strategy. Different from the traditional scout bee strategy, we use a local search strategy when a bee stagnates for a defined value. The experimental results on Berkeley segmentation database demonstrate the effectiveness of our algorithm.
Xingyu Xia, Hao Gao, Haidong Hu, Rushi Lan, Chi-Man Pun

Dynamic Consolidation Based on Kth-Order Markov Model for Virtual Machines

Abstract
The rapid development of cloud computing technology has led to a high level of energy consumption. The central processing unit (CPU) of the data center and other resources often use less than half the rate; therefore, if the work of the virtual machine is focused on part of the server, and the idle server switches to low power mode, the power consumption of the data center can be greatly reduced. Traditional research into virtual machine consolidation is mainly based on the high load threshold of the current host load setting or periodically migrates, and the present study made predictions based on the timing of problems of lower prediction accuracy faced. To solve these problems, we consider the impact of the multi-order Markov model and the CPU state at different times, and propose a new hybrid sequence K Markov model for the next period of time of the host CPU load forecasting. Owing to the large-scale data experiment on the CloudSim simulation platform, the host load forecasting method proposed in this paper is compared with the traditional load detection method to verify that the proposed model has a large reduction in the number of virtual machine migrations and amount of data center energy consumption, and the violation of the service level agreement (SLA) is also at an acceptable level.
Na Jiang

Research into the Adaptability Evaluation of the Remote Sensing Image Fusion Method Based on Nearest-Neighbor Diffusion Pan Sharpening

Abstract
Nearest-neighbor diffusion pan sharpening, as a new image fusion method based on nearest-neighbor diffusion, has become a new hot spot of research. In this paper, the nearest-neighbor diffusion pan sharpening method is used for a WorldView-2 image fusion experiment and compared with the methods we usually use such as the wavelet transform fusion method, the PCA transform fusion method, and the Gram–Schmidt transform fusion method. The experimental results show that the spatial information is better than the other three methods in terms of spatial details and texture.
Chunyang Wang, Weikuan Shao, Huimin Lu, Hebing Zhang, Shuangting Wang, Handong Yue

Estimation of Impervious Surface Distribution by Linear Spectral Mixture Analysis: A Case Study in Nantong, China

Abstract
In recent years, with rapid expansion of cities, natural ecological landscapes centering on green environments such as vegetation have been gradually replaced by impervious buildings. Consequently, a severe influence that cannot be ignored has been imposed on the whole ecological environment. In this paper, the main urban area of Nantong of China is used as a study area. Landsat 8 satellite remote-sensing images are used as a data source and linear spectral unmixing method is utilized to extract impervious surface information of the city and to study the distribution conditions of impervious surface percentage (ISP). The experimental analysis indicates the closer to the commercial area and highly intensive residential area, the bigger the ISP will become.
Ping Duan, Jia Li, Xiu Lu, Cheng Feng

Marine Organisms Tracking and Recognizing Using YOLO

Abstract
A system that investigates deep sea automatically has never developed. A purpose of this study is developing such a system. We employed a technique of recognition and tracking of multi-objects, called “You Only Look Once: YOLO.” This method provides us very fast and accurate tracker. In our system, we remove the haze, which is caused by turbidity of water, from image. After its process, we apply “YOLO” to tracking and recognizing the marine organisms, which includes shrimp, squid, crab, and shark. Our developed system shows generally satisfactory performance.
Tomoki Uemura, Huimin Lu, Hyoungseop Kim

Group Recommendation Robotics Based on External Social-Trust Networks

Abstract
Recommendation robotics helps users to find similar interests or purposes to those of others. We often provide advice to close friends or similar users, such as sharing favorite dishes, listening to favorite music, etc. In traditional group recommendation robotics, however, users’ personalities have been ignored. In this chapter, a method of group recommendation robotics based on social-trust networks is proposed, which builds a group profile by analyzing not only users’ preferences, but also the social relationships between members inside and outside of the group. We employ a collaborative filter to obtain members’ predictions and adjust the final group preference rating by the external social-trust network if the group has a large disagreement. The experimental results show that the proposed method has a lower root mean square error and leads to a satisfactory effect for the group.
Guang Fang, Lei Su, Di Jiang, Liping Wu

Vehicle Logo Detection Based on Modified YOLOv2

Abstract
Vehicle logo detection technology is one of the research directions in the application of intelligent transportation systems. It is an important extension of detection technology based on license plates and motorcycle types. A vehicle logo is characterized by uniqueness, conspicuousness, and diversity. Therefore, thorough research is important in theory and application. Although numerous vehicle logo detection methods exist, most of them cannot achieve real-time detection for different scenes. The YOLOv2 network is improved by constructing the data of a vehicle logo, dimension clustering of the bounding box, reconstructing network pre-training, and multi-scale detection training. This work implements fast and accurate vehicle logo detection. The generalization of the detection model and anti-interference capability in real scenes are optimized by data enrichment. The experimental results show that the accuracy and speed of the detection algorithm are improved.
Shuo Yang, Chunjuan Bo, Junxing Zhang, Meng Wang

Energy-Efficient Virtual Machines Dynamic Integration for Robotics

Abstract
The rapid development of cloud computing technology has brought a lot of energy consumption. However, the utilization rate of resources such as data center CPUs is often less than half. Therefore, if the virtual machines in operation are centrally integrated into some servers, and idle servers are switched to low-power modes, the power consumption of data centers can be greatly reduced. The consumption. The traditional research on the integration of virtual machines is mainly based on the current load of the host to set a high-load threshold or periodically perform the migration. At present, research based on time-series prediction faces the problem of low prediction accuracy. In order to solve these problems, this paper synthetically considers the influence of multi-order Markov model and CPU state at different times, and proposes a new K-order mixed Markov model for CPU load prediction of the host for a period of time in the future. By conducting large-scale data experiments on the CloudSim simulation platform, the host load forecasting method proposed in this paper is compared with traditional load detection methods, and the proposed model is greatly reduced in the number of virtual machine migrations and data center energy consumption. And the violation of the SLA is also at an acceptable level.
Haoyu Wen, Sheng Zhou, Zie Wang, Ranran Wang, Jianmin Lu

Multi-Level Chaotic Maps for 3D Textured Model Encryption

Abstract
With the rapid progress of virtual reality and augmented reality technologies, 3D contents are the next widespread media in many applications. Thus, the protection of 3D models is primarily important. Encryption of 3D models is essential to maintain confidentiality. Previous work on encryption of 3D surface model often considers the point clouds, the meshes, and the textures individually. In this work, a multi-level chaotic maps model for 3D textured encryption was presented by observing the different contributions for recognizing cipher 3D models between vertices (point cloud), polygons, and textures. For vertices which make main contribution for recognizing, we use high-level 3D Lu chaotic map to encrypt them. For polygons and textures which make relatively smaller contributions for recognizing, we use 2D Arnold’s cat map and 1D logistic map to encrypt them, respectively. The experimental results show that our method can get similar performance with the other method and use the same high-level chaotic map for point cloud, polygons, and textures, while we use less time. Besides, our method can resist more method of attacks such as statistic attack, brute-force attack, and correlation attack.
Xin Jin, Shuyun Zhu, Le Wu, Geng Zhao, Xiaodong Li, Quan Zhou, Huimin Lu

Blind Face Retrieval for Mobile Users

Abstract
Recently, cloud storage and processing have been widely adopted. Mobile users in one family or one team may automatically backup their photos to the same shared cloud storage space. The powerful face detector trained and provided by a 3rd party may be used to retrieve the photo collection which contains a specific group of persons from the cloud storage server. However, the privacy of the mobile users may be leaked to the cloud server providers. In the meanwhile, the copyright of the face detector should be protected. Thus, in this paper, we propose a protocol of privacy preserving face retrieval in the cloud for mobile users, which protects the user photos and the face detector simultaneously. The cloud server only provides the resources of storage and computing and cannot learn anything from the user photos and the face detector. We test our protocol inside several families and classes. The experimental results reveal that our protocol can successfully retrieve the proper photos from the cloud server and protect the user photos and the face detector.
Xin Jin, Shiming Ge, Chenggen Song, Le Wu, Hongbo Sun

Near-Duplicate Video Cleansing Method Based on Locality Sensitive Hashing and the Sorted Neighborhood Method

Abstract
With the wide utilization of intelligent video surveillance technology, increasing amounts of near-duplicate video has been generated, which seriously affects the data quality of the video data set. Cleaning this dirty data automatically from the video data set has become an important issue that needs to be urgently resolved. In this chapter, a near-duplicate video cleansing method based on locality sensitive hashing (LSH) and the sorted neighborhood method (SNM) is presented in an attempt to solve the above problem. First, the speeded-up robust feature is extracted from the video and then the sorted candidate set is built by using LSH; on this basis, the near-duplicate videos are cleaned by using the SNM. Finally, the simulation experiments are implemented to show that the presented method in this chapter is effective, which can be used to clean near-duplicate videos automatically and improve video data quality.
Ou Ye, Zhanli Li, Yun Zhang

A Double Auction VM Migration Approach

Abstract
Virtualization technology plays an important role in cloud computing. Virtual machine (VM) migration can reduce the cost of cloud computing data centers. In this paper, a double auction-based VM migration algorithm is proposed, which takes the cost of communication between VMs into account under normal operation situation. The algorithm of VM migration is divided into two parts: (1) selecting the VMs to be migrated according to the communication and occupied resources factors of VMs, (2) determining the destination host for VMs which to be migrated. We proposed VMs greedy selection algorithm (VMs-GSA) and VM migration double auction mechanism (VMM-DAM) to select VMs and obtain the mappings between VMs and underutilized hosts. Compared with other existing works, the algorithms we proposed have advantages.
Jinjin Wang, Yonglong Zhang, Junwu Zhu, Yi Jiang

An Auction-Based Task Allocation Algorithm in Heterogeneous Multi-Robot System

Abstract
Today, robots are facing dynamic, real-time, complex adversarial, and stochastic work environment. It’s of great significance to research on task allocation problem in multi-robot system. In this paper, we propose a dynamic auction method for differentiated tasks under cost rigidities (DAMCR) which can find the optimal result in a static auction between robots and tasks. Then we analyze the optimality of DAMCR. Considering the dynamics of the system, we propose a partition-based task allocation adjustment method via distributed approach. To verify the effectiveness of the proposed scheme, we compare it with other task allocation methods based on classic Hungarian algorithm. In the experiments, we analyze the impact of the number of tasks and robots on the running time and robustness of the methods. The results suggest that our solution outperforms others, that is, robots can accomplish tasks faster and more effectively.
Jieke Shi, Zhou Yang, Junwu Zhu

Non-uniformity Detection Method Based on Space-Time Autoregressive

Abstract
The inhomogeneous phenomena of nonhomogeneity of clutter power, interference target and isolated interference are always coexisting in the real environment of airborne radar. Therefore research on new inhomogeneous detection methods applied to the case of coexisting several inhomogeneous phenomena has become an important subject in the field of research on radar signal detection technology. The new combined space-time autoregressive (STAR) algorithm is proposed for suppressing all three kinds of inhomogeneous phenomena, while the existing STAR algorithms have no capacity, and the proposed algorithm can suppress all three kinds of inhomogeneous phenomena effectively that is indicated in the results of simulation. The simulation results show the effectiveness of the proposed algorithm.
Ying Lu

Secondary Filter Keyframes Extraction Algorithm Based on Adaptive Top-K

Abstract
As the coal mine environment is similar to night-time, there is less discernible information, which makes the coal mine video images collected by the camera have a high level of redundancy, less available information, obvious light spots, and noise interference, which are not conducive to extracting useful information from the video. In view of the above problems, a keyframes extraction algorithm for coal mine video images based on a secondary filter with adaptive Top-K is proposed. The algorithm calculates the eigenvalues of the feature points using the principal component analysis method, then filters the eigenvalues by the threshold of adaptive Top-K to extract the effective keyframes of the coal mine image. The experimental results show that the algorithm can extract the keyframes more accurately using the adaptive threshold method.
Yan Fu, Chunlin Xu, Mei Wang

An Introduction to Formation Control of UAV with Vicon System

Abstract
As a first step towards fully distributed formation control for a large swarm, we exhibit a distributed flight demo with three quadcopters under Vicon motion capture system. In this paper, hardware setup of system and a decentralized software architecture are introduced. Then, in order to validate the system, we design an experiment in which three quadcopters circle around a fixed point coordinately by using the consensus protocol. Finally, some flight data are reported and analyzed.
Yangguang Yu, Zhihong Liu, Xiangke Wang

Quadratic Discriminant Analysis Metric Learning Based on Feature Augmentation for Person Re-Identification

Abstract
The quadratic discriminant analysis (XQDA) method learns a general projection matrix for all cameras with its strong generalization ability, but it ignores the inherent properties of each camera itself and does not take feature of changes in each camera into account, causing each person under the camera to have a certain feature distortion problem which makes its discriminative ability worse. In this paper, feature augmentation is used to enhance the inherent properties of each camera. By ensuring the generalization ability of the camera, the feature of changes within each camera is taken into consideration and the final discriminative ability is improved. Finally, experiments on a challenging person re-identification dataset, VIPeR, show that the proposed method outperforms the state-of-the-art methods.
Cailing Wang, Hao Qi, Guangwei Gao, Xiaoyuan Jing

Weighted Linear Multiple Kernel Learning for Saliency Detection

Abstract
This paper presents a novel saliency detection method based on weighted linear multiple kernel learning (WLMKL), which is able to adaptively combine different contrast measurements in a supervised manner. Three commonly used bottom-up visual saliency operations are first introduced, including corner-surround contrast (CSC), center-surround contrast (CESC), and global contrast (GC). Then these contrast measures are fed into our WLMKL framework to produce the final saliency map. We show that the assigned weights for each contrast feature maps are always normalized in our WLMKL formulation. In addition, the proposed approach benefits from the advantages of the contribution of each individual contrast operation, and thus produces more robust and accurate saliency maps. The extensive experimental results show the effectiveness of the proposed model, and demonstrate the combination is superior to individual subcomponent.
Quan Zhou, Jinwen Wu, Yawen Fan, Suofei Zhang, Xiaofu Wu, Baoyu Zheng, Xin Jin, Huimin Lu, Longin Jan Latecki

Backmatter

Weitere Informationen