AI-Powered ADAS for Traffic Light Detection: Improving Driver Safety, Inclusivity and Traffic Efficiency

2025-28-0086

To be published on 02/07/2025

Event
Advances in Design, Materials, Manufacturing and Surface Engineering for Mobility (ADMMS’25)
Authors Abstract
Content
Recent advancements in artificial intelligence, particularly in machine learning, natural language processing, and computer vision, have catalyzed innovation across diverse industries. The evolution of Advanced Driver Assistance Systems (ADAS) in contemporary mobility has markedly improved driver safety and convenience. This paper introduces a novel AI-driven ADAS technology engineered to detect and interpret traffic light signals, providing real-time assistance to drivers. This system is particularly advantageous for individuals with colour vision deficiencies and aims to enhance traffic efficiency. The system is based on an AI-powered device equipped with a high-resolution camera and advanced image recognition algorithms to accurately classify traffic light colours. The device conveys this information to the driver through both visual and auditory feedback mechanisms and is designed for seamless integration into the dashboard of any modern vehicle. The AI model leverages transfer learning, utilizing the YOLOv8 algorithm trained on the LISA dataset, which encompasses US road conditions, supplemented by additional datasets from various Indian cities. This enables robust performance in diverse traffic environments. The ADAS device performs edge inference, eliminating reliance on cloud-based GPUs, thereby minimizing latency and ensuring real-time feedback. By providing precise and clear signals about traffic light statuses, this system enhances driving confidence and safety for colour-blind drivers, promoting inclusivity. Additionally, real-time alerts on traffic light changes improve overall traffic efficiency by reducing driver reaction times and mitigating traffic congestion through the prevention of traffic shockwaves. Moreover, the data collected by the system serves as a foundational platform for enhancing interactions between traditional traffic lights and autonomous vehicles, fostering a harmonious coexistence of conventional and autonomous vehicles. This AI-driven ADAS innovation represents a significant advancement in automotive technology, aiming to make driving safer, more inclusive, and more efficient. It contributes to smoother traffic flow in urban environments and lays the groundwork for autonomous vehicles.
Meta TagsDetails
Citation
Tom, J., "AI-Powered ADAS for Traffic Light Detection: Improving Driver Safety, Inclusivity and Traffic Efficiency," SAE Technical Paper 2025-28-0086, 2025, .
Additional Details
Publisher
Published
To be published on Feb 7, 2025
Product Code
2025-28-0086
Content Type
Technical Paper
Language
English