Advance Obstacle Detection for Autonomous Vehicles Using Numerical Data from LIDAR and RADAR Sensor: A Machine Learning Approach

Abstract: Obstacle detection is a critical aspect of autonomous vehicle systems, ensuring safe navigation in dynamic environments. This paper presents a novel approach to improving obstacle detection by integrating LIDAR and RADAR data using a robust sensor fusion framework, along with advanced clustering algorithms such as DBSCAN. The proposed method enhances the accuracy and reliability of obstacle identification by reducing noise through advanced preprocessing techniques and optimizing the clustering parameters. Extensive testing in various real-world scenarios demonstrated that the algorithm significantly reduces false positives while improving precision and recall metrics. In addition, the system achieves real-time performance, ensuring immediate responses in challenging driving conditions. Compared to existing methods, the proposed approach provides superior accuracy, adaptability to different environments, and a more comprehensive obstacle detection process. This work contributes to the development of safer and more reliable autonomous vehicles. Keywords: Autonomous vehicles, Obstacle detection, Sensor fusion, LIDAR, RADAR, Clustering algorithms. Title: Advance Obstacle Detection for Autonomous Vehicles Using Numerical Data from LIDAR and RADAR Sensor: A Machine Learning Approach Author: Rabaya Akter, Kaniz Fatema Oyshee, Md Nagib Mahfuz Sunny, PROMANANDA ROY, Faysal Ahammed, Md Faysal Refat International Journal of Novel Research in Engineering and Science ISSN 2394-7349 Vol. 11, Issue 2, September 2024 - February 2025 Page No: 20-29 Novelty Journals Website: www.noveltyjournals.com Published Date: 19-October-2024 DOI: https://doi.org/10.5281/zenodo.13955333 Paper Download Link (Source) https://www.noveltyjournals.com/upload/paper/Advance%20Obstacle%20Detection%20for%20Autonomous-19102024-7.pdf