This paper presents a point-based 3D object detection framework for urban driving. It integrates LiDAR range filtering and region proposals to detect vehicles and obstacles with higher accuracy in traffic-heavy conditions.

Isabella Quinn, Robert Hastings, Emily Dawson, Ethan Wallace, Jessica Harper

This research enhances visual scene interpretation in challenging weather. Using synthetic data and multi-domain generalization, the system maintains semantic segmentation reliability during snow, fog, and heavy rain in autonomous vehicle deployments.

Chen Hao Ming, Liu Wen Sheng, Zhang Liang Zhi, Gao Tian Wei, Xu Rong Jie

This paper proposes a recurrent network for visual odometry in GPS-denied zones. It fuses motion-aware flow with learned depth sequences, allowing accurate navigation in tunnels, dense cities, and subterranean infrastructures.

Abigail White, Logan Pierce, Hannah Brooks, Marcus Doyle, Grace Thompson

This study presents a VQA model for embodied AI. It uses adaptive encoding and object grounding to support navigation through question-based reasoning in unfamiliar environments, enabling intelligent interaction with surroundings.

Chen Ming Zhi, Liu Fang Hao, Zhang Rui Hui, Xu Jian Bo, Gao Ping Liang

This study presents an underground SLAM system that combines inertial sensing with visual self-supervision. It ensures reliable localization in GPS-denied subterranean settings, supporting search-and-rescue and industrial tunnel inspection tasks for autonomous ground robots.

Chen Yu Bo, Liu Wen Jie, Xu Hao Rui, Zhang Min Tao, Gao Fang Cheng

This paper proposes an attention-enhanced U-Net architecture for aerial scene parsing. By fusing multi-altitude data, the model improves accuracy in terrain mapping and obstacle detection, enabling precise drone navigation in cluttered environments.

Rebecca Hall, Noah Edwards, Madison Turner, Elijah Hamilton, Sophie Mitchell

This research develops a terrain classification framework for planetary rovers. Using contrastive domain adaptation and stereo vision, it improves traction control and path planning across unfamiliar and rough terrain types.

Chen Min Hao, Liu Bo Chen, Xu Zhi Fang, Zhang Wen Yang, Gao Tian Zhi