Design and Evaluation of a New Machine Learning Framework for IoT and Embedded Devices

• DOI: 10.3390/electronics10050600
• Publication Date: 2021-03-05T05:39:07Z
• Authors: Gianluca Cornetta; Abdellah Touhafi
• Keywords: Leverage (statistics)

Low-cost, high-performance embedded devices are proliferating and a plethora of new platforms available on the market. Some them either have GPUs or possibility to be connected external Machine Learning (ML) algorithm hardware accelerators. These enhanced features enable applications in which AI-powered smart objects can effectively pervasively run real-time distributed ML algorithms, shifting part raw data analysis processing from cloud edge device itself. In such context, Artificial Intelligence (AI) considered as backbone next generation Internet Things (IoT) devices, will no longer merely collectors forwarders, but really “smart” with built-in wrangling that leverage lightweight machine learning algorithms make autonomous decisions field. This work thoroughly reviews analyses most popular particular emphasis those more suitable resource-constrained devices. addition, several been built top custom multi-dimensional array library. The designed framework has evaluated its performance stressed Raspberry Pi III- IV-embedded computers.