Revolutionary Enhanced Lane Departure Detection Techniques for Autonomous Vehicle Safety using ADAS

Abstract

This paper introduces an Enhanced Lane Departure Warning System (ELDWS) leveraging cutting-edge vision technology to boost road safety for ADAS enabled vehicles. Our novel approach utilizes a combination of Phase Stretch Transform for edge detection, Curved and Straight Lane Detectors for accurate lane Detection, and a modified Kalman Filter for dynamic lane tracking, collectively aimed at improving vehicle safety through precise lane departure estimation. Unlike traditional systems that rely heavily on clear lane markings and favorable environmental conditions, our model excels in various lighting and road scenarios, including curved paths and challenging weather conditions. The research demonstrates the system's efficacy in real-world simulations, where it outperforms existing technologies in detecting and alerting potential lane departures. Through meticulous integration of advanced image processing techniques and machine learning algorithms, our model offers a significant leap towards achieving robust lane-keeping assistance in autonomous and semi-autonomous vehicles. Furthermore, the paper discusses the system's ability to adapt to different environmental conditions and road types, making it a versatile tool for enhancing driving safety. By addressing the limitations of current LDWS technologies, such as sensitivity to weather conditions and the reliance on high-contrast lane markings, our approach sets a new standard for safety in the autonomous driving domain. This paper shows that by overcoming the problems of older lane departure warning systems, like their struggle with bad weather and dependency on clear road markings, our system sets a new standard for keeping cars safely within their lanes, especially in self-driving and semi-self-driving cars. We offer a detailed solution that combines new detection methods with advanced tracking techniques for better accuracy and reliability. The accuracy of our ELDWS is quantitatively measured using a custom dataset, revealing impressive results. During daytime conditions, our system achieved a high accuracy rate of 95.81%, correctly detecting lanes in 7665 out of 8000 frames. This demonstrates the system's robustness in optimal lighting conditions. The accuracy remains commendable in left and right departure scenarios during the day, with rates of 83.26% and 84.83%, respectively, showcasing the system's capability to recognize lane departures effectively. This research is important for making cars safer and provides useful information for developing better driving assistance systems in the future.