A Comprehensive Taxonomy for AIoT Video Streaming Systems: Components, Connectivity, and Applications

The fusion of Artificial Intelligence (AI) technologies with the Internet of Things (IoT) has given rise to the innovative paradigm known as AIoT (Artificial Intelligence of Things). In this context, AIoT video streaming emerges as a vital application, leveraging the potential of AI to process and analyze video data from IoT devices in real-time. To facilitate a better understanding of the complex ecosystem surrounding AIoT video streaming, this paper presents a comprehensive taxonomy. Our taxonomy encompasses various aspects of AIoT video streaming, including the underlying devices, connectivity protocols, data processing and analytics techniques, security and privacy considerations, video streaming methods, and diverse applications across industries. We explore the roles of IoT devices and edge computing in capturing and processing video data, while addressing the significance of cloud connectivity for scalable and centralized analytics. The taxonomy further delves into the implementation of AI algorithms for video analysis, enabling tasks such as object recognition, facial detection, and anomaly detection. Security and privacy challenges are of utmost importance in AIoT video streaming, and our paper highlights the significance of data encryption, access control, and compliance with regulations. Furthermore, we discuss various real-time streaming protocols, adaptive bitrate streaming, and video compression techniques that optimize data transmission and storage. In conclusion, this taxonomy serves as a comprehensive guide for researchers, developers, and industry professionals in understanding the intricacies of AIoT video streaming systems. By providing a systematic categorization of the components and applications, this paper lays the groundwork for future advancements in AIoT video streaming, enabling the development of intelligent and scalable video analytics solutions across various domains