This paper introduces an adversarial domain randomization technique. Realism-guided perturbations and consistency-aware training improve synthetic-to-real generalization, enabling robust simulation-to-reality transfer for self-driving model validation and development workflows.

Anirudh Jayanth Ramesh, Violet Hannah Monroe, Sun Hao Jie, Minato Takahiro Oda, Annabelle Kayleigh Pierce

This research proposes a gesture-based interaction system for autonomous robots. Cross-view keypoints are matched with action templates using temporal decoding, allowing seamless interpretation of human commands in dynamic, vision-driven collaborative environments.

Yash Vikram Singh, Natalie Hope Carmichael, Liang Zhao Hui, Kenichi Hideo Morita, Sierra Madeline Vaughan

This work presents a SLAM system optimized for micro drones. Sparse depth filters and adaptive keyframe selection reduce energy demands while maintaining localization accuracy, enabling extended flight time for autonomous aerial mapping missions.

Rohit Surya Talreja, Hannah Joy Underwood, Zhang Wei Bin, Kojiro Haru Yamamoto, Emily Rose Hartman

This paper introduces a visual anomaly detection system. Uncertainty is measured through entropy and distribution modeling in feature space, enabling robots to flag unreliable perception outputs during autonomous exploration or task execution.

Tarun Aditya Mehta, Violet Paige Saunders, Huang Bo Xing, Takumi Hiroshi Nakagawa, Penelope Isla Graham

This work presents a multiview normal estimation framework. By combining cross-angle visual features and hybrid reflectance priors, it achieves accurate surface orientation prediction under dynamic or non-uniform lighting environments encountered by mobile robots.

Deepak Harish Nair, Felicity Rae Middleton, Wu Qing Shan, Noboru Issei Taniguchi, Tessa Brynn Franklin

This paper introduces a graph SLAM system for semi-structured indoor environments. Adaptive loop closure detection based on keyframe similarity enables robust topological mapping for long-term autonomous indoor navigation applications.

Ramesh Devinder Chauhan, Ivy Mae Whitaker, Sun Liang Jie, Sho Takashi Hayashi, Margaret Claire Redding