Real time Alert AI and IoT based accident prevention and detection
Open Access
- 1 February 2024
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 275-286
Abstract
Road incidents caused by driver distractions (such as phone usage), drunkenness, and excessive speed constitute significant dangers to public safety in Rwanda. A comprehensive study of the design and implementation of an AI and IoT-based system for accident detection, prevention, and enhanced emergency response is presented in this article as one of the solutions to the problem. Using Raspberry Pi microcontroller as the central server, the system incorporates several sensors and actuators connected to ESP8266, including GPS Gn801 GPS Glonass dual mode for speed detection and car localization, MQ3 alcohol gas sensor SEN42, R16 for alcohol detection, ADXL335-3 Axis compass accelerometer GY-6l for car movement changes. In addition, a Raspberry Pi 4Bl3B 5MP fisheye night vision focal camera is integrated into the system to detect driver distractions such as phone usage while driving. Sophisticated machine learning and artificial intelligence (AI) algorithms, with a specific focus on Convolutional Neural Networks (CNNs), were utilized to identify instances of driver distraction and facilitate real-time monitoring and alert systems. By leveraging on edge and cloud computing to process, store and analyze data, the system guarantees efficient and impactful correspondence with emergency response teams. Our objective is to contribute to the domain of road safety by developing an efficient solution for emergency response and accident prevention in Rwanda. Our primary focus is on our distinctive contribution, which is to address driver distraction monitoring and real-time reporting to a person in charge in hospital and police in an innovative manner. This addresses a crucial aspect that is currently not sufficiently addressed by other solutions in the field. The study's findings demonstrate the accuracy of the integrated IoT sensors in accurately detecting accidents, speed, monitoring alcohol consumption, and identifying distractions, specifically phone usage. A remarkable 86-100% accuracy is attained by Custom Convolutional Neural Networks (CNNs) when it comes to identifying driver distractions. The combination of edge and cloud computing guarantees timely and effective communication with emergency response teams. This study highlights the critical role that technology plays in improving road safety by proposing a technologically advanced solution for emergency response and accident prevention in Rwanda. The capabilities of the system to monitor in real-time and its efficacy in accident detection and prevention demonstrate its potential to make a substantial impact on the field of road safety.Keywords
This publication has 15 references indexed in Scilit:
- A Machine Learning Framework for Automated Accident Detection Based on Multimodal Sensors in CarsSensors, 2022
- Vehicle Accident Detection & Alert System using IoT and Artificial Intelligence2021
- Deep neural network-based predictive modeling of road accidentsNeural Computing and Applications, 2020
- IoT based framework for the detection of vehicle accidentCluster Computing, 2019
- Deep learning in big data Analytics: A comparative studyComputers and Electrical Engineering, 2019
- Wireless vehicle alert and collision prevention system design using Atmel microcontroller2016
- GPS and Map matching based vehicle accident detection system2013
- A System for Automatic Notification and Severity Estimation of Automotive AccidentsIEEE Transactions on Mobile Computing, 2013
- Automatic Accident Detection: Assistance Through Communication Technologies and VehiclesIEEE Vehicular Technology Magazine, 2012
- Mobile phone location determination and its impact on intelligent transportation systemsIEEE Transactions on Intelligent Transportation Systems, 2000