International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 11 Issue: 05 | May 2024
www.irjet.net
p-ISSN: 2395-0072
ARTIFICIAL INTELLIGENCE BASED ACCIDENT PREVENTION SYSTEM S. Thirumalairajan1, B. Dharanitharan2, S. Dwaraga Prasath3, S. Gugan4, V. Logeshwaran5 1Lecturer, Department of Mechatronics, PSG Polytechnic College, Coimbatore-4, Tamil Nadu, India
2345 Student Scholar, Department of Mechatronics, PSG Polytechnic College, Coimbatore-4, Tamil Nadu, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract – The report describes the creation of a complex
intricate process is concealed from users, who can benefit from requiring less data and training time to develop effective and accurate models. PictoBlox, on the other hand, is a graphical programming software modeled after the latest version of Scratch, designed to make coding an enjoyable and straightforward experience. Featuring a user-friendly interface and drag-and-drop functionality, it serves as an ideal platform for beginners venturing into the realm of programming.
system aimed at tracking head gestures and preventing accidents by using machine learning, graphical programming, and hardware components. By employing a webcam, the system detects head movements and analyzes them using a trained machine learning model. It then triggers responses through hardware devices controlled by an ESP8266 microcontroller. The report begins by introducing Teachable Machine and PictoBlox, highlighting their roles in creating machine learning models without coding. It then explains the stages of the machine learning process, from defining objectives to deploying models. Additionally, it reviews existing AI-based accident prevention systems, discusses hardware integration, and explores the functions of specific components such as the ESP8266, buzzer, and vibrator motor. Detailed explanations of hardware components, interfacing with PictoBlox, and example programs are provided, along with step-by-step instructions for block programming in PictoBlox, including machine learning integration. Finally, the report concludes with abstract, program, and output sections summarizing the project's goals, methodologies, and achievements, demonstrating the effective use of machine learning, graphical programming, and hardware integration to enhance safety proactively.
2. LITERATURE REVIEW Numerous approaches have been explored in prior studies to tackle distracted driving incidents, incorporating both technological advancements and behavioral interventions aimed at reducing driver distractions and enhancing road safety.
Key Words: Accident Prevention, Artificial intelligence, Machine Learning, ESP8266, Teachable Machine, Safety, Buzzer, Vibrator Motor, Pictoblox, Block Program.
Cognitive Distraction Assessment: Previous research has examined various methods for assessing cognitive distraction among drivers, including eye tracking, cognitive workload assessments, and reaction time tests. By monitoring these factors, researchers can identify periods of heightened distraction and develop targeted interventions to mitigate their impact.
Behavioral Interventions: Researchers have also explored behavioral interventions to address distracted driving behaviors. These interventions encompass educational campaigns, enforcement of distracted driving laws, and the implementation of incentive-based programs to promote safe driving practices. By focusing on driver attitudes and behaviors, these interventions aim to cultivate a culture of distraction-free driving.
Technological Solutions: Technological advancements have led to the development of innovative solutions to combat distracted driving. These solutions include smartphone applications that disable certain functions while driving, invehicle systems that offer real-time feedback on driver behavior, and vehicle-to-infrastructure communication systems that alert drivers to potential hazards. By harnessing technology, these solutions strive to mitigate the impact of distractions on driving performance.
1.INTRODUCTION Teachable Machine is an online tool designed for individuals to train their own machine learning classification models without any coding knowledge. Users can utilize their webcam, images, or sound to train models. It employs transfer learning, a machine learning technique, to identify patterns and trends within the provided data and swiftly generate a classification model. Through transfer learning, users can incorporate their own data and refine a model based on a pre-existing base model that has been trained on a vast dataset within a specific domain. For instance, the base model (mobilenet) used for image recognition in Teachable Machine was initially trained to recognize 1000 classes (such as dogs, phones, beds, trombones, etc.). The inherent features utilized by mobilenet to recognize these classes can also be applied to identify new classes defined by the user. This
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 1422