Search

The Internet of Things (IoT) is an important component of the new digital infrastructure and is deeply integrated with the fifth-generation mobile communication (5G), big data, cloud computing, artificial intelligence (AI), blockchain, and digital twin. It is profoundly changing the technology system and promoting the digital economy, ushering in a new stage of smart IoT system in which everything is connected. This paper reviews the development status of IoT in China, proposes the concept of smart IoT system (IoT 2.0), and expounds on the implications, architecture, technical pedigree, and key enabling technologies. Practical cases of smart IoT system are explored considering the application scenarios of intelligent manufacturing, smart agriculture, smart grid, smart healthcare, intelligent transportation, and intelligent environmental protection, demonstrating the application values of the smart IoT system. Furthermore, we suggest that a technology integration innovation project that integrates IoT, AI, 5G, and new application field technologies should be implemented; focus should be placed on the research, development, and industrialization of intelligent products such as smart IoT systems / cloud native platforms / low-code (no-code) application development environments and toolsets, high-end sensors for smart IoT systems, and IoT chips / special components; and application demonstration of cloud-edge-end collaborative, autonomous controllable, safe, and credible smart IoT systems should be conducted.

The new intelligent manufacturing system integrates information and communication technologies (ICTs) with industrial technologies and supports the rapidly unfolding new round of industrial revolution. The rapid evolution of ICTs gives intelligent manufacturing the potential for accelerated development. This paper proposes a technical system for the new intelligent manufacturing system and elaborates the connotation and characteristics of a technical subsystem regarding ICTs from the perspective of industrial Internet system technologies and four basic technologies, namely, industrial big data, artificial intelligence (AI), fifth-generation mobile communication (5G), and modeling simulation/digital twin. Subsequently, we present the vertical, horizontal, and end-to-end application scenarios of intelligent manufacturing enabled by ICTs, and propose several suggestions for promoting the new intelligent manufacturing system through ICTs. First, special science and technology projects should be established focusing on advanced networks, collaborative computing, and industrial knowledge reasoning. Second, Industrial development should focus on the R&D and industrialization of the following technologies: 5G application, network collaboration, intelligent and intelligently connected products based on new-generation AI technology, and domestication of modeling simulation/digital twin tool sets and systems. Third, Application demonstration should be conducted regarding industrial design based on 5G Plus industrial virtual reality, industrial platforms for AI Internet of Things, and intelligent design of industrial products based on modeling simulation/digital twin technology. Meanwhile, it is necessary to improve the efficient and collaborative working mechanism for promoting new intelligent manufacturing policies, accelerate the construction of interconnection standard groups, promote the industrial chain and supply chain through industry–university–research collaboration, and enhance the deep integration of industry and education in intelligent manufacturing.

Aiming at the problem that the isolation of trust relationship between different security domains of cloud manufacturing system leads to the repealed identity authentication of users, trust domain access and the rejection of cloud service cross domain collaboration, an inter domain mutual trust process model for cloud manufacturing system is deigned. An efficient trusted security optimization technology of user authentication and service trust domain collaboration based on inter domain mutual trust is proposed, which realizes the trust domain transmission of users, trusted identity and cloud manufacturing service cross domain collaboration. The application is verified in some enterprises, and a comparative analysis with the traditional method is given. The analysis results show that the efficient and trusted security technology proposed can improve the efficiency of cloud manufacturing system authentication and service trust domain collaboration without redoing the protection intensity of the existing security mechanism. [ABSTRACT FROM AUTHOR]

Manifold learning is a promising intelligent data analysis method, and the manifold learning preserves the local embedding features of the data in manifold mapping space. Manifold learning has its limitations on extracting the nonlinear features of the data in many applications. For example, hyperspectral image classification needs to seek the nonlinear local relationships between spectral curves. For that, researchers applied the kernel trick to manifold learning in the previous works. The kernel-based manifold learning was developed, but still endures the problem that the inappropriate kernel model reduces the system performance. In order to solve the problem of kernel model selection, we propose a manifold framework of multiple-kernel learning for the application of hyperspectral image classification. In this framework, the quasiconformal mapping-based multiple-kernel model is optimized based on the optimization objective equation, which maximizes the class discriminant ability of data. Accordingly, the discriminative structure of data distribution is achieved for classification with the quasiconformal mapping-based multiple-kernel model. [ABSTRACT FROM AUTHOR]

Kernel machine is a feasible and effective nonlinear feature extraction method on data analysis, for example, hyperspectral sensing data. Kernel trick improves largely the performance of learning system including recognition, clustering, prediction through the nonlinear kernel mapping from the input data space to output data space. The performance of kernel-based system is largely influenced by the function and parameter of kernel. Optimizing only the parameters is not effective to promote the kernel-based learning system, because the data distribution is not changed only changing the kernel parameter. Moreover, no any universal single kernel is very adaptive to all applications. We present a framework of quasiconformal mapping-based kernel learning machine under single kernel and multiple kernels for hyperspectral image data classification. The performance of learning system is improved largely owing to the two facts: quasiconformal kernel structure changes the data structure in the kernel empirical space; quasiconformal multikernels characterize precisely the data for improving performance on solving complex visual learning tasks. The learning framework is applied to the hyperspectral image classification, and some experiments are implemented on two hyperspectral image databases. [ABSTRACT FROM AUTHOR]