Coordination and Control of Ground-Aerial Robotic Systems

本研究计划的目标是为包含无人地面车辆（UAV）和微型飞行器（MAV）的异构多机器人系统的协同操作开发控制和定位技术。最近，许多行业已经开始探索近距离监测环境和关键基础设施的其他可能性。MAVs可以在人类无法进入的区域进行导航，并能够提供环境的3D视角。有效载荷限制禁止MAVs在更大范围内自主部署更长时间，也引入了物理限制，无法携带先进的传感器套装和强大的计算机进行精确定位。有两个主要问题需要解决。首先，无人机通常部署在受限的地区，而且在大多数情况下，GPS无法访问。在这些情况下，MAV的机载传感器定位能力差，不足以实现准确定位。第二个问题是，传感器的测量范围很短，这使得它们难以覆盖更大的区域。大多数轨迹控制系统依赖于外部跟踪系统的反馈；因此，这些控制器对感应体积之外的操作无效。另一方面，ugv具有更高的续航能力，可以进行更长的操作，更快的处理，远距离传感和高度精确的定位信息。因此，探索任务正逐步向多智能体实现方向发展，以利用它们的互补能力。在这种协作方法中，MAVs可以使用地面机器人的相对测量来引导，计算要求高的任务可以委托给强大的ugv。因此，本研究的重点是开发使用ugv的MAVs辅助导航和控制系统。在短期内，该计划试图解决以下关键领域的几个研究主题；(a)基于机器人间感知的MAV轨迹跟踪控制，(b)基于cubature Kalman滤波方法的MAV相对定位，(c)基于模糊离散事件系统和模型预测控制的混合控制，以及(d)多机器人分布式控制和探索。本提案拟提供4个博士级和2个硕士级HQP培训。加拿大的许多航空航天和海洋行业正在尝试在许多海底和北极勘探任务中使用自动驾驶汽车。本方案解决的本地化和控制问题非常适用于此类应用。因此，该项目将产生与加拿大许多工业部门相关的HQP培训和高影响力研究。

The objective of this research program is to develop control and localization technology for cooperative operation of a heterogeneous multi-robotic system containing both unmanned ground vehicles (UAV) and micro-aerial vehicles (MAV). Recently many industries have begun to explore alternative possibilities of monitoring the environment and critical infrastructures at close distances. The MAVs can navigate through areas that are inaccessible for humans and have the ability to provide 3D perspective of the environment. The payload constraints prohibit MAVs to be deployed autonomously for longer durations covering a wider area and also introduce physical limitations to carry advance sensor suits and powerful computers for accurate positioning. There are two major issues to be resolved. First is that drones are typically deployed in constrained areas and in most cases where GPS is not accessible. The poor localization available with on-board sensors of an MAV is not sufficient for accurate positioning in these settings. The second issue is that the sensors have short range of measurements which makes them difficult to be deployed covering a wider area. Most trajectory control systems rely on feedback obtained from external tracking systems; as a result these controllers are invalid for operations outside the sensing volume. On the other hand, UGVs are available with higher endurance for longer operations, faster processing, long range sensing, and highly accurate localization information. Therefore, exploration missions are progressively moving towards multi-agent implementations to harness their complementary capabilities. In this collaborative approach, the MAVs can be guided using relative measurements taken by the ground robots and the computationally demanding tasks can be delegated to the powerful UGVs. Therefore, this research focuses on developing aided navigation and control systems for MAVs using UGVs. In the short term this program attempts to address several research themes in the following key areas; (a) trajectory tracking control of MAVs using inter-robot sensing, (b) MAV relative localization using cubature Kalman filtering approaches, (c) hybrid control using fuzzy discrete event system and model predictive control and (d) multi-robotic distributed control and exploration. This proposal expects to provide four doctoral level and two Masters level HQP training . Many aerospace and ocean industries in Canada are now attempting to use autonomous vehicles for many sub-sea and arctic exploration missions. The localization and control problems addressed in this proposal are highly applicable for such applications. Therefore, the project will produce HQP training and high impact research relevant across many industrial sectors in Canada.