This study presents a real-time autonomous navigation model integrating vision and LiDAR data using deep learning. The system enhances environmental awareness and obstacle detection while reducing processing latency, supporting safe and sustainable deployment of self-driving vehicles in complex urban road networks.

Rahul Santhosh Iyer, Jean Claude Bourget, Priya Meenakshi Ramakrishnan, Fatima Noor Al-Hussein, Carlos Miguel Duarte

This paper presents a semantic segmentation network optimized for autonomous systems in dynamic cityscapes. The model leverages attention mechanisms and real-time adaptive learning to enhance precision in object boundary detection, enabling safer navigation in traffic-dense metropolitan scenarios.

Chen Rui Bo, Liu Hao Zhen, Xu Fang Min, Zhang Wen Liang, Gao Ming Sheng

This study introduces a hybrid approach for depth estimation from single-camera inputs. By combining transformer-based fusion layers and sparse stereo constraints, the model produces precise scene depth maps, enhancing safety and spatial reasoning in autonomous ground vehicle systems.

Olivia Reed, Jonathan Meyers, Emily Rhodes, Benjamin Lawson, Sophia Clarke

This paper proposes a novel low-light object detection framework tailored for nighttime self-driving conditions. Using noise-resilient convolutional layers and illumination-adjusted feature normalization, the system accurately detects pedestrians, vehicles, and hazards under poor lighting conditions.

Rajat Kumar Sinha, Fatima Noor Al-Bahrani, Jean Marcel Rousseau, Priya Lakshmi Venkat, Carlos Rafael Silva

This research develops a real-time facial recognition system for in-vehicle alertness detection. Cross-domain transfer learning boosts generalization across lighting and facial variations, while lightweight CNNs reduce processing load on embedded devices.

Chen Yu Wei, Xu Bo Zhi, Liu Jie Rong, Zhang Tian Cheng, Gao Wen Xuan

This study proposes a combined point cloud registration and semantic labeling pipeline for indoor robots. It enables real-time 3D mapping and path planning for autonomous service robots navigating within dynamic and cluttered interior spaces.

Isabelle Clarke, Henry Woods, Megan Doyle, Owen Bradford, Charlotte Howard

This paper develops a predictive model that combines attention mechanisms with real-time motion vectors to estimate pedestrian intent. It enhances decision-making accuracy at intersections, improving the safety of autonomous vehicle operations in dense urban areas.

Chen Ming Tao, Liu Tian Zhi, Xu Hao Long, Zhang Wen Rui, Gao Fang Ming

This study proposes a visual-inertial odometry system for autonomous vehicles operating in GPS-restricted zones. By recalibrating feature confidence and employing lightweight sensor fusion, the system maintains trajectory estimation accuracy with low computational overhead.

Sarah Bennett, Michael Harper, Daniel Thornton, Abigail Foster, Christopher Lambert