ABSTRACT
Typical architectures designed for the Internet of Things (IoT) are adopted for intelligent-based networks. The adopted architectures are based on software-defined networks, cloud, edge-based architectures, and mobility-based architectures. However, Artificial Internet of Things (AIoT) networks need more scalable, flexible, energy-efficient, and interoperable-based architectures where the systems can integrate a huge number of devices. AI-empowered IoT architecture needs merged cloud, fog, and edge systems tailored with deep learning capabilities. This chapter discusses the architectures, techniques, and existing platforms for AIoT networks. This chapter also presents a layered-based architecture specially designed for AIoT networks based on a cloud computing layer, edge layer, and fog layer. The Edge layer provides all functions of the perception layer like smart sensing, visual sensing, and monitoring services, and can control the device's perception and behavior. The fog layer can control devices such as routers, switches, and gateways. The last cloud layer is responsible for handling network applications, functions, services, and intelligent integration for the network. The proposed architecture supports and provides all the required services of the AIoT networks.
