| 摘要: |
| 多智能体的协同视觉SLAM系统在地面机器人或空中飞行器梯队中应用得越来越广泛。由于不同的智能体之间会交换或共享信息，协同视觉SLAM系统不同于应用在单个智能体上的典型的视觉SLAM系统，具有更好的鲁棒性、准确性和高效性。阐述了多智能体协同视觉SLAM的发展历史和现有的研究方法，描述了协同视觉SLAM系统实现所涉及的关键技术，包括协同姿态估计和地图构建，以及新兴的分布式架构。此外，对多智能体协同视觉SLAM技术的发展趋势进行了展望。 |
| 关键词: 多智能体 协同 视觉SLAM 地图构建 |
| DOI： |
|
| 基金项目:国家重点研发计划（2016YFB0501600）；装备预研基金重点项目（6140517020101） |
|
| Research Progress in Collaborative Visual SLAM for Multiple Agents |
| WANG Lu,YANG Gong-liu,CAI Qing-zhong,CHEN Wei,YAN Xu-liang |
| (School of Instrumentation and Optoelectronic Engineering, Beihang University, Beijing 100083, China) |
| Abstract: |
| The collaborative visual SLAM system of multiple agents is more and more widely used in ground robot or air vehicle echelon. Because different agents exchange or share information, collaborative visual SLAM system is different from the typical visual SLAM system applied to a single agent. The collaborative visual SLAM system has better robustness, accuracy and efficiency. This paper describes the development history and existing research methods of multi-agent collaborative visual SLAM, and describes the key technologies involved in the realization of collaborative visual SLAM system, including collaborative attitude estimation and mapping, as well as the emerging distributed architecture. In addition, the development trend of multi-agent collaborative visual SLAM technology is prospected. |
| Key words: Multiple agents Collaborative Visual SLAM Map Construction |
