Machine Learning on Microcontrollers: A Review of TinyML Frameworks and Prospects

TinyML denotes the deployment of machine learning models on very small, low-power devices. This review summarises a survey of the TinyML field and the description of the TensorFlow Lite Micro framework, covering the motivation, the design constraints of microcontroller targets and the challenges identified in these sources. This is a narrative review; it contains no new data.