面向无人机避碰的可行且计算高效路径规划
机器人学
2021-02-09 v2
摘要
本文提出了一种鲁棒且计算高效的无人机（Unmanned Aerial Vehicle, UAV）实时避碰算法，即基于记忆的沿墙-人工势场（Memory-based Wall Following-Artificial Potential Field, MWF-APF）方法。该新算法在沿墙法（Wall-Following Method, WFM）与人工势场法（Artificial Potential Field method, APF）之间切换，并具备改进的态势感知能力。历史轨迹被纳入考虑以避免重复的错误决策。此外，该方法可有效应用于计算能力较低的平台。作为示例，采用配备有限数量飞行时间（Time-of-Flight, TOF）测距仪的四旋翼来验证该算法的有效性与效率。通过软件仿真与实物飞行测试，证明了 MWF-APF 方法在复杂场景中的能力。
引用
@article{arxiv.2002.10623,
title = {Feasible Computationally Efficient Path Planning for UAV Collision Avoidance},
author = {Han Wang and Muqing Cao and Hao Jiang and Lihua Xie},
journal= {arXiv preprint arXiv:2002.10623},
year = {2021}
}
备注
IEEE International Conference on Control and Automation (ICCA) 2018