Implementing AI on microcontrollers in fog and edge architectures
Dalibor Dobrilović · 2023
We are witnessing the era of evolution of complex system architectures, starting from cloud computing and gradually to the edge, fog, and mist computing paradigms. The development of applications in those complex systems is equally important for companies as well as for universities. Both branches need fast, low-cost, and effective environments for rapid prototyping. These demands can be enhanced with the recent microcontroller developments. The microcontrollers grow in their performance and, simultaneously, are available for considerably lower prices. Open-source hardware movement is another important base for rapid fog and edge prototyping. This paper presents the approach of using open-source hardware development boards in edge computing architectures by investigating the possibility of implementing Artificial Intelligence on these edge devices. The performance of AI-supported edge devices is monitored and analyzed in terms of their speed, capacity, and accuracy. The Arduino, ESP8266, and ESP32 development boards are evaluated in this research.