Search

One typical fog-enabled intelligent IoT system is the smart home, where each smart appliance/device is able to connect to the Internet and carry out some computing tasks. Each appliance/device can be viewed as an IoT node. These IoT nodes form a local network. To enable the home to better understand the humans and subsequently respond correctly, an efficient and secure human machine interact technology is necessary. Conventional remote controls are extremely inconvenient due to the larger number of appliances and the dependence on the hardware. A more efficient solution is to let the local network itself recognize the user behavior directly. Radio-based behavior recognition has advantages in smart home scenarios where comforts and privacy protection are of our major concern. Meanwhile, numerous wireless communications between the IoT nodes in the smart home also facilitate the implementation of these approaches. In this chapter, we will mainly focus on this type of behavior recognition. Besides, we can also take advantage of the acoustical signals to track the moving objects. Specifically, the speakers and microphones in cell phones can be employed to transmit and receive the sound signals. As accurate user behavior recognition becomes possible due to fog computing, our homes will surely become smarter and smarter.

Internet of Things (IoT) is a technology that aims at providing connectivity for anything, by embedding short-range mobile transceivers into a wide array of additional gadgets and everyday items, enabling new forms of communication between people and things, and between things themselves. This chapter reviews key IoT technologies and several applications, which include not only simple data sensing, collection, and representation, but also complex information extraction and behavior analysis. As 5G mobile networks are beginning to be commercially deployed worldwide, intelligent IoT applications and services are getting more and more important and popular in different business sectors and industrial domains, thanks to more communication bandwidth, better data quality, faster data rate, denser network connectivity, lower transmission latency, and higher system reliability.

Robots are used widely in many fields now, such as earthquake rescue, smart factory, and so on. They bring us lots of convenience in daily lives, save huge manpower in factories, and help to complete many mission-possible tasks in some cases. For these applications, simultaneous localization and mapping (SLAM), efficient management, and collaboration among robots are necessary. However, in these robot applications, it may suffer issues of high cost, large power consumption, and low efficiency. An effective solution is to employ fog computing. This chapter introduces fog-enabled solutions for robot SLAM, multi-robot smart factory, and multi-robot fleet formation applications, which require large local computing power for timely constructing the map of a working environment, calculating multiple robots’ exact positions, and tracking their movement postures and orientations. Through a high-speed wireless network, massive data and images collected by onboard and local sensors are transmitted from the robots and intelligent infrastructure to nearby fog nodes, where intelligent data processing algorithms are responsible for analyzing valuable information and deriving the results in real time.

Intelligent transportation system (ITS) helps to improve traffic efficiency and ensure traffic safety. The core of this system is the collection and analysis of sensor data and vehicle communication technologies. The challenges of ITS mainly focus on two aspects: computing and communication, while security and interoperability are the prerequisites of the system. Existing network architecture and communication technology still cannot meet the demand for advanced intelligent driving support and rapid development of intelligent transportation. As an emerging concept, fog computing is proposed for various IoT scenarios and can address the challenges in intelligent transportation system. Fog computing enables the critical functions of ITS by collaborating, cooperating, and utilizing the resources of underlying infrastructures within roads, smart highways, and smart cities. Fog computing will address the technical challenges in ITS and will help scale the deployment environment for billions of personal and commercial smart vehicles. In this chapter, we first introduced the definition and development of ITS, describing the ecosystem composition and their respective requirements. Then, we explained the challenges and a stage-of-the-art of ITS, mainly focusing on vehicle station and communication network. To present fog computing, the architecture of fog-enabled ITS was provided. And we also discussed how fog computing can address the technical challenges and provide strong support for ITS. Finally, several use cases in fog-enabled ITS, including autonomous driving, cooperative driving, and shared vehicles, are shown in this chapter, which further verifies the benefits that fog computing can bring to ITS.

Fog computing is a horizontal, system-level architecture that distributes computing, storage, control, and networking functions closer to the users along a cloud-to-thing continuum. This chapter introduces the architecture and key enabling technologies of fog computing, as well as its latest development in standardization bodies and industrial consortium. As the bridge connecting the cloud and things, fog computing plays the crucial role in identifying, integrating, managing, and utilizing multi-tier computing, communication, and storage resources in different IoT systems. Together with feasible AI algorithms, fog nodes can combine various local/regional micro-services and orchestrate more intelligent applications and services with different user preferences and stringent performance requirements. For example, autonomous driving and intelligent manufacturing require high security in data transmission and storage, very low latency in data processing and decision making, and super-high reliability in network connectivity and service provisioning. Further, the challenges of developing more sophisticated services across multiple domains are discussed.

For many fog-enabled application scenarios, such as multi-robot systems, wireless communication networks, intelligent transportation systems, and smart home, they can be generally modeled as Multi-Task Multi-Helper (MTMH) fog networks. Specifically, consider a general heterogeneous fog network consisting of different types of Fog Nodes (FNs), wherein some Task Nodes (TNs) have computation-intensive and delay-sensitive tasks, while some Helper Nodes (HNs) have spare computation resources for sharing with their neighboring nodes. How to effectively map multiple tasks or TNs into multiple HNs to minimize every task’s delay in a distributed manner is a fundamental challenge, which is called the MTMH problem. This chapter proposes an analytical framework for general MTMH fog networks. To be specific, a comprehensive system model consisting of network architecture, wireless channels, communication and computing models, and task types is developed for a MTMH fog network. Based on different game theories, the fundamental problems of computation offloading are formulated and analyzed for non-splittable and splittable tasks, respectively. Accordingly, two efficient algorithms are designed and fully evaluated under different performance metrics.

Wireless communication networks are experiencing an unprecedented traffic growth and an increasing variety of services, each with potentially different traffic patterns and quality of service (QoS) and/or quality of experience (QoE) requirements. To cope with the continuing traffic growth and service expanding, future wireless networks will have to be heterogeneous and densely deployed, featuring the coexistence of different radio access technologies (RATs), and will be significantly more complex to deploy and operate than the existing wireless networks. This has made it evident for the necessity of wireless network self-optimization, where wireless networks are automated to minimize human intervention and to proactively optimize network deployment, operation, and multi-RAT resource allocation to meet increasing service demand from people and the Internet of Things (IoT). Recently, fog computing has been considered as a promising paradigm shift to enable autonomous management and operation of wireless networks. Since research on fog-enabled wireless network self-optimization has just started, there are many aspects that are not well understood and many open challenges that need to be addressed. In this chapter, we explore how fog computing would enable self-optimization for wireless networks, which will act as the infrastructure to provision ubiquitous wireless connectivity for the IoT. More specifically, we will first discuss different self-organizing network (SON) architectures and how they would benefit from the fog computing paradigm, and then look into how fog computing would provide new opportunities and enable new features for several important SON functionalities, including mobility load balancing, self-optimization of mobility robustness and handover, self-coordination of inter-cell interference, self-optimization of coverage and capacity, and self-optimized allocation of computing, storage, and networking resources in wireless networks.