Robust Autonomous Landing of Multirotor Unmanned Aerial Vehicles on Static and Moving Platforms

自主无人机因其灵活性和快速部署而在各个行业越来越受欢迎，并已在搜救、检查和测绘等应用中显示出实用性。我们在自主无人机操作方面的新研究方向是稳健地自主降落在静止和在风中移动的平台上，使无人机能够与地面车辆(卡车)或船舶集成。这将在递送服务、农业、搜救、检查、边境管制和海上行动方面带来科学突破和实践创新。目前，无人机不具备在足够小的着陆区重复着陆的能力，无法实现强大的卡车或舰载行动。该项目的目标是开发新的建模、状态估计和控制方法，以实现无人机在风中的精确和一致着陆，最终实现自动着陆在移动平台上，如卡车货床和船舶。该项目由四个研究支柱组成：(1)空气动力学建模，以建立多旋翼无人机飞行的准确数学模型，特别是在湍流条件下；(2)状态估计，以能够估计飞行过程中对无人机的风力影响；(3)运动规划，以自动找到着陆在移动平台上的最佳路径，同时考虑风的影响和着陆平台的任意运动，以及(4)飞行控制，以确保无人机对风况和着陆平台运动的快速响应，并密切跟踪无人机成功着陆的最优着陆路径。通过真实世界的飞行实验来评估算法的有效性。该项目将帮助加拿大保持其在这一领域的关键作用，并将导致短途运输方面的创新。

Autonomous drones are becoming more and more popular in various industry sectors for their flexibility andfast deployment and have demonstrated usefulness in applications such as search and rescue, inspection, andmapping. Our new research direction in autonomous drone operations is robust autonomous landing on staticand moving platforms in wind to enable the integration of drones with ground vehicles (trucks) or ships. Thiswill lead to scientific breakthroughs and practical innovations in delivery services, agriculture, search andrescue, inspections, border control, and maritime operations.Presently, drones do not have the capability to land repeatably in a small-enough landing zone to enable robusttruck- or ship-based operations. This project's goal is to develop new modeling, state estimation, and controlmethods for the precise and consistent landing of drones particularly in wind, ultimately enabling automatedlanding on moving platforms such as truck cargo beds and ships.This project consists of four research pillars: (1) aerodynamic modeling to develop accurate mathematicalmodels of multi-rotor drone flight, especially in turbulent wind conditions, (2) state estimation to enableestimation of wind effects on the drone during flight, (3) motion planning to automatically find the optimalpath to land on a moving platform while considering wind effects and arbitrary motion of the landing platform,and (4) flight control to ensure drones' fast response to wind conditions and motion of landing platform, andclose following of the optimal landing path for successful drone landing.The effectiveness of the algorithms will be evaluated through real-world flight experiments. Thisproject will help Canada to maintain its key role in this sector and will lead to innovation in short-hop delivery.