# Bubble Explorer: Fast UAV Exploration in Large-Scale and Cluttered 3D-Environments using Occlusion-Free Spheres

**Authors:** Benxu Tang, Yunfan Ren, Fangcheng Zhu, Rui He, Siqi Liang, Fanze Kong, Fu Zhang  
**arXiv:** [2304.00852](https://arxiv.org/abs/2304.00852)  
**Date:** 2023-04-03  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2304.00852v1/

## Key points

- Autonomous exploration is crucial in robotics applications
- Recent advancements aim to generate global coverage paths for enhanced exploration efficiency
- Bubble Explorer is a novel method designed for fast UAV exploration in large-scale and cluttered 3D environments
- Bubble Explorer utilizes occlusion-free spheres to generate viewpoints efficiently
- The approach combines greedy strategy and global optimization for a balance between efficiency and effectiveness
- Extensive benchmarking shows superior performance in exploration efficiency and computational time
- Real-world experiments validate the efficacy of Bubble Explorer in navigating complex terrains with minimal computational resources

## Summary

In the field of robotics, autonomous exploration plays a vital role in various applications. Traditional methods have focused on immediate rewards for goal selection, but may not lead to optimal overall exploration efficiency. Recent advancements in state-of-the-art methods aim to generate global coverage paths, significantly enhancing exploration efficiency. However, these approaches often come with high computational overhead and infrequent planner updates. To address these challenges, researchers including Benxu Tang, Yunfan Ren, Fangcheng Zhu, Rui He, Siqi Liang, Fanze Kong and Fu Zhang have introduced Bubble Explorer - a novel method designed for fast UAV exploration in large-scale and cluttered 3D environments. This innovative approach utilizes occlusion-free spheres to generate viewpoints in a computationally efficient manner. One unique aspect of Bubble Explorer is its combination of both greedy strategy and global optimization to strike a balance between computational efficiency and exploration effectiveness. Extensive benchmarking against existing methods has shown superior performance in terms of exploration efficiency and computational time. Real-world experiments further validate the efficacy of this approach in navigating complex terrains while minimizing computational resources. Overall, Bubble Explorer represents a significant advancement in autonomous exploration for UAVs by leveraging occlusion-free spheres and combining different optimization strategies for enhanced efficiency in challenging environments.