3.8 Proceedings Paper

A data-stream TinyML compression algorithm for vehicular applications: a case study

出版社

IEEE
DOI: 10.1109/MetroInd4.0IoT54413.2022.9831606

关键词

Internet of Things; Online data Compression; TinyML; Embedded Algorithm

资金

  1. Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior -Brasil (CAPES) [001]
  2. Brazilian fostering agency CNPq (National Council for Scientific and Technological Development) [435683/2018-7]

向作者/读者索取更多资源

The Internet of Things (IoT) has brought challenges of potentially irrelevant or redundant data transmission, and increased data creation and transmission speed. To address these challenges, data compression techniques and TinyML can be used to implement machine learning models on low-power devices.
The Internet of Things (IoT) has reduced the distance between one point and another and between people by connecting multiple devices to the web. However, the volume and speed of data creation and transmission have also increased. In this scenario, some challenges start to emerge, such as potentially irrelevant or redundant data transmission, that is, generating a more significant expenditure of energy and processing, in addition to the unnecessary use of the communication channel. Thus, to mitigate these, a solution for IoT devices would be data compression techniques. However, such devices available in the market today have severe limitations in terms of storage and processing power. Therefore, to overcome these limitations, the TinyML can be used to seek ways to implement machine learning models in low-power devices. In this context, this article aims to evaluate the impact of the compression algorithm (Tiny Anomaly Compress - TAC) on the performance of a microcontroller applied to the context of vehicles in a real scenario. As a result, it was found that even with the embedded algorithm, the microcontroller processing time is not affected in a meaningful way.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

3.8
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

暂无数据
暂无数据