Navigation of unmanned vehicles for civil applications within confined spaces

无人驾驶地面车辆（UGV）已被用于自然，恐怖主义和意外任务，有时成功，但更多的是证实其潜力。加拿大皇家骑警、加拿大搜救队（SAR）和军方已经考虑在行动中使用UGV。UGV在9/11事件中表现不错，但没有取得巨大成功，因为机器人无法在倒塌的建筑物内执行任务。UGV在已知地形中表现出相对的成功，但由于许多原因而没有真实的成功，但主要是由于它们不能在未知的复杂障碍的室内和室外城市环境中移动，并且不能有效地操纵21/2 D和3D地形。在以前工作的基础上，我们将开发今天还没有的导航工具，当与现有工具相结合时，将使UGV能够在SAR，侦察和城市警务领域执行任务，使UGV能够在有限的空间内高速自主移动。目标应用程序侧重于关键业务，旨在支持民防机构，消防队，EMS等，我们的目标是两个科学领域：导航和控制。这项工作将侧重于行为，目标和策略选择，以及探索机制，这些机制将在SAR，警察和监视场景中进行测试。导航区域最初将侧重于2½D和3D导航运动。未来五年的阶段包括：i）为在有限的3D空间中工作的UGV开发导航系统; ii）为高速UGV在未知的2.5 D和3D地形中导航开发地形评估方法; iii）将步骤（i）和（ii）合并为一个新的工具，考虑到机器人的能力和地形条件。步骤（i）将根据机器人的范围/视觉信息及其运动/重新配置能力开发机器人运动工具。地形的3D地图将用于开发动态配置空间，机器人将从该空间确定克服地形条件的运动。步骤（ii）会发展一个计算成本较低的机制，以根据货车的参数和地形轮廓/特徴的变化，估计货车的速度和转弯角。

Unmanned Ground Vehicles (UGVs) have been used in natural, terrorists and accidental missions sometimes with success, but more often just confirming their potential. The RCMP, Canada Search & Rescue (SAR) and the military have considered using UGVs for their operations. UGVs did well, during the 9/11 incident but without a huge success as the robots couldn't perform tasks within the collapsed buildings. UGVs have shown relative success in known terrains but no real success otherwise due to a number of reasons but mainly due to their inability to move within unknown complex obstructed indoor and outdoor urban environments and effectively maneuver 2½D & 3D terrains. Based on previous work we will develop navigational tools not available today that when coupled with existing tools will enable UGVs execute missions in the areas of SAR, reconnaissance and urban policing enabling UGVs move autonomously within confined spaces at high speeds. The target applications focus on critical operations aimed at supporting civil protection agencies, fire brigades, EMS, etc. We target two scientific areas: navigation and control. The work will focus on behaviors, goal & strategy selection, and exploration mechanisms that will be tested on SAR, police and surveillance scenarios. The navigation area will initially focus on 2½D and 3D navigation motion. The phases for the next five years include: i) Develop a navigation system for UGVs working in confined 3D spaces; ii) Develop a terrain assessment method for high speed UGVs navigating in unknown 2½D & 3D terrains; and iii) Merge Steps (i) and (ii) into a novel tool taking into consideration the robot's capabilities and the terrain conditions. Step (i) will develop robot motion tools based on the robot's range / visual information and its motion / reconfiguration capabilities. A 3D map of the terrain will be used to develop a dynamic configuration space from where the robot will determine the motions to overcome the terrain conditions. Step (ii) will develop a computationally inexpensive mechanism to estimate the velocity and turn angle of the UGV based on the UGV's parameters and changes in the terrain profile/characteristics.