Reproduce “MMAUD: A Comprehensive Multi-Modal Anti-UAV Dataset for Modern Miniature Drone Threats” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Shenghai Yuan, Yizhuo Yang, Thien Hoang Nguyen, Thien-Minh Nguyen, Jianfei Yang, Fen Liu, Jianping Li, Han Wang and 1 more

The MMAUD dataset from Nanyang Technological University offers a comprehensive multi-modal resource for anti-UAV research, integrating stereo vision, Lidars, Radar, and an audio array. It provides millimeter-level accurate 3D ground truth for drone detection, classification, and precise 3D trajectory estimation, demonstrating strong 2D visual detection (YoloX achieving 85.9% mAP) and highlighting challenges for audio-based localization in noisy environments.

Key insights

  • The integration of diverse sensors (stereo vision, Lidars, Radar, audio array) provides a richer data source for robust anti-UAV tasks than single-modality datasets.
  • Millimeter-level accurate 3D ground truth, generated by a Leica MultiStation, offers unprecedented precision for evaluating anti-UAV algorithms.
  • Designing the dataset with cost-effective sensors and open-source accessibility aims to foster broader adoption and development of deployable, mobile-ready anti-UAV solutions.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.