基于视觉的多无人机协同探索系统研究  

Research on Vision-based Multi-UAV Collaborative Exploration System

作　　者：章坤 李瑞[1] 陈聪 史莹晶[1] ZHANG Kun;LI Rui;CHEN Cong;SHI Yingjing(School of Automation Engineering,University of Electronic Science and Technology of China,Chengdu 611731,China)

机构地区：[1]电子科技大学自动化工程学院,四川成都611731

出　　处：《无线电工程》2025年第3期621-632,共12页Radio Engineering

基　　金：国家自然科学基金(61973055);四川省自然科学基金(2023NSFSC0511)。

摘　　要：针对当前无人机探索方案普遍存在的效率低、光线投射过程耗时较长以及规划算法效率不足等问题,创新性地提出了一种基于视点球和拓扑图的高效视觉探索系统。该系统优化了探索策略,并设计了多机通信策略及多机视点选取策略,将单无人机探索系统拓展为分布式多无人机协同探索系统,有效提高了在未知环境中的探索效率。实际实验数据表明,提出的基于视觉的多无人机协同探索系统,比现有主流方案在平均探索时间上缩短了6.5%,在平均计算时间上缩短了77%,对无人机在复杂、未知环境下进行高效自主探索与协同工作具有重要意义。To address the prevalent issues in current UAV exploration schemes,such as low efficiency,lengthy light projection processes,and insufficient efficiency of planning algorithms,an efficient vision-based exploration system based on viewpoint spheres and topological maps is innovatively proposed.This system optimizes exploration strategies and designs multi-UAV communication strategies and multi-UAV viewpoint selection strategies,extending the single-UAV exploration system into a distributed multi-UAV collaborative exploration system.This effectively enhances exploration efficiency in unknown environments.Actual experimental data demonstrates that the proposed vision-based multi-UAV collaborative exploration system reduces the average exploration time by 6.5%and the average computation time by 77%compared to existing mainstream solutions,which is significant for efficient autonomous exploration and collaborative work of UAVs in complex and unknown environments.

关 键 词：自主探索 规划 建图 无人机 分布式系统 

分 类 号：TP391.7[自动化与计算机技术—计算机应用技术]