A TinyML-based Alcohol Impairment Detection System For Vehicle Accident Prevention

计算机工程与设计 Pub Date : 2022-09-23 DOI:10.1109/SEEDA-CECNSM57760.2022.9932962

Anargyros Gkogkidis, Vasileios Tsoukas, A. Kakarountas

{"title":"A TinyML-based Alcohol Impairment Detection System For Vehicle Accident Prevention","authors":"Anargyros Gkogkidis, Vasileios Tsoukas, A. Kakarountas","doi":"10.1109/SEEDA-CECNSM57760.2022.9932962","DOIUrl":null,"url":null,"abstract":"Driving under the influence of alcohol is one of the most severe and critical problems in every country throughout the world. Driving is a difficult endeavor that demands a high degree of concentration and great visual processing. A system based on the Internet of Things can be utilized to measure drivers’ alcohol level and restrict their operation of motor vehicles. This technology is affordable but has a number of disadvantages, including the requirement for an internet connection, the transfer of data to other organizations, bandwidth and latency constraints, and security concerns. TinyML is an emerging technology that can overcome the aforementioned challenges by performing machine learning models locally and delivering real-time intelligence. In this work, the possibility of developing a TinyML-based system that can detect alcohol and alert the driver was investigated. The experimental findings demonstrate a high degree of accuracy, indicating that the technology under consideration may be utilized to develop compact, intelligent, and inexpensive devices capable of detecting alcohol and alerting the driver in real-time.","PeriodicalId":68279,"journal":{"name":"计算机工程与设计","volume":"24 1","pages":"1-6"},"PeriodicalIF":0.0000,"publicationDate":"2022-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"计算机工程与设计","FirstCategoryId":"1093","ListUrlMain":"https://doi.org/10.1109/SEEDA-CECNSM57760.2022.9932962","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 1

Abstract

Driving under the influence of alcohol is one of the most severe and critical problems in every country throughout the world. Driving is a difficult endeavor that demands a high degree of concentration and great visual processing. A system based on the Internet of Things can be utilized to measure drivers’ alcohol level and restrict their operation of motor vehicles. This technology is affordable but has a number of disadvantages, including the requirement for an internet connection, the transfer of data to other organizations, bandwidth and latency constraints, and security concerns. TinyML is an emerging technology that can overcome the aforementioned challenges by performing machine learning models locally and delivering real-time intelligence. In this work, the possibility of developing a TinyML-based system that can detect alcohol and alert the driver was investigated. The experimental findings demonstrate a high degree of accuracy, indicating that the technology under consideration may be utilized to develop compact, intelligent, and inexpensive devices capable of detecting alcohol and alerting the driver in real-time.

查看原文本刊更多论文

基于tinyml的汽车事故预防酒精损伤检测系统

酒后驾车是世界上每个国家最严重和最关键的问题之一。驾驶是一项困难的工作，需要高度集中的注意力和出色的视觉处理能力。基于物联网的系统可以用来测量驾驶员的酒精含量，并限制他们驾驶机动车。这种技术经济实惠，但有许多缺点，包括需要互联网连接、向其他组织传输数据、带宽和延迟限制以及安全问题。TinyML是一项新兴技术，可以通过在本地执行机器学习模型并提供实时智能来克服上述挑战。在这项工作中，研究了开发一种基于tinyml的系统的可能性，该系统可以检测酒精并提醒驾驶员。实验结果显示了高度的准确性，表明正在考虑的技术可用于开发紧凑、智能和廉价的设备，能够检测酒精并实时提醒驾驶员。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

计算机工程与设计

自引率

0.00%

发文量

20353

期刊介绍： Computer Engineering and Design is supervised by China Aerospace Science and Industry Corporation and sponsored by the 706th Institute of the Second Academy of China Aerospace Science and Industry Corporation. It was founded in 1980. The purpose of the journal is to disseminate new technologies and promote academic exchanges. Since its inception, it has adhered to the principle of combining depth and breadth, theory and application, and focused on reporting cutting-edge and hot computer technologies. The journal accepts academic papers with innovative and independent academic insights, including papers on fund projects, award-winning research papers, outstanding papers at academic conferences, doctoral and master's theses, etc.