Using Microcontrollers for the Inference of ML Algorithms

IAP 2020 - 2021
As time passes, the number of Internet-of-things (IoT) applications keeps growing exponentially, leading to a large amount of data that needs to be processed daily and with low latency [1]. Consequently, a new computing model is required to optimize network performance and support all the volume and variations of data. One of such architectures, known as ​ Edge Computing ​ , requires a main hardware design solution capable of scaling with the amount of data while also maintaining a consistent throughput. Field Programmable Gate Arrays (FPGA) are a viable solution that satisfies the requirements for such applications. With this in mind, the Embedded Systems Research Group (ESRG) wishes to take part in the development of edge computing technology and its applications. The group's main focus is to learn and comprehend different microcontrollers, using them as classifiers for machine learning algorithms. Then incorporate the skills acquired and be able to participate in more in-depth projects that apply edge computing to the data gathered from a trained algorithm.