Reproduce “Fast-Tracker: A Robust Aerial System for Tracking Agile Target in Cluttered Environments” 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
Zhichao Han, Ruibin Zhang, Neng Pan, Chao Xu, Fei Gao
Fast-Tracker, an aerial system developed by Zhejiang University, enables robust and safe autonomous tracking of agile targets in cluttered environments by integrating constrained B ezier regression for target motion prediction and a hierarchical, spatial-temporal optimal trajectory planner. The system achieved a 15Hz replanning frequency and demonstrated superior prediction accuracy (e.g., 1.82m error vs. 3.70m for baseline in low noise) and computational efficiency (e.g., 7.3ms total computation vs. 234.0ms for baseline) in real-world and simulated scenarios.
Key insights
- Constrained B ezier regression can provide more accurate and robust target motion predictions, especially under noisy sensor data, by enforcing dynamic feasibility through control point constraints.
- A hierarchical trajectory planning approach, integrating a kinodynamic search with spatial-temporal optimization, effectively balances safety, dynamic feasibility, and computational efficiency for high-frequency replanning.
- Actively re-locating the target through a combined path and trajectory generation strategy significantly enhances the overall reliability and robustness of aerial tracking systems in complex environments.
- Orienting the quadrotor tangentially to its tracking trajectory using a multi-camera setup can improve obstacle observation and facilitate safe navigation in cluttered surroundings.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.