CONTROL OF SELF-ORGANIZING FORMATIONS OF AUTONOMOUS VEHICLES

许多工业和服务业务需要具有监视和数据收集能力的任务。这适用于室外远程和城市环境，城市以及建筑物和其他室内设施。这些任务可以由自动驾驶车辆的集中编队执行。在这种情况下，中央控制站获取所需的所有数据，制定控制命令并将其传送到所有单元，并且这些操作可能会受到传感器信号和控制命令的通信故障的影响。 这种应用的更好解决方案是自动驾驶车辆的自组织编队。它们需要新的路径规划和控制策略。 在该建议中，考虑了两种代表性情况，即小尺寸和大尺寸地层。根据不同的使命，一组自动驾驶车辆可能必须作为一个大型编队运行，而在其他任务中，只需要减少数量的车辆，并将作为一个小型编队运行。 对于小尺寸编队，所提出的方法是基于预定义的几何编队形状之间的切换，以提供在自由场和狭窄通道的环境中的平滑操作。几何编队形状将从历史上使用的经过验证的实际形状中选择，例如在军事行动中：排，线性波前，V形编队，矩形等，每一个都对特定的预定环境有效。 对于大尺寸地层，实际地层形状将由地层保持要求和特定环境约束的并发影响产生。这种自组织编队方法将不基于预定义的几何形状，并且适合于时变数量的自主车辆的混合。 本项目中考虑的自动驾驶车辆具有本地感测和本地驱动，并且在大多数情况下，感测不足和驱动不足。

Numerous industrial and service operations require missions with surveillance and data collection capabilities. This applies to outdoor remote and in urban environments, in cities as well as in buildings and other indoor facilities. These missions can be performed by centralized formations of autonomous vehicles. In that case a central control station acquires all data needed, elaborates control commands and communicates them to all units and these operations can be subject to communication failures of sensors signals and control commands. Better solutions for such applications are self-organizing formations of autonomous vehicles. They would require new path planning and control strategies. In this proposal two representative cases are considered, small size and large size formations. Depending on the mission, a group of autonomous vehicles might have to operate as a single large formation, while in other missions only a reduced number of vehicles are needed and will operate as a small formation. For small size formations the proposed approach is based on switching among predefined geometric formation shapes to provide smooth operation in environments with free fields and narrow passages. Geometric formations shapes will be selected from proven practical shapes used historically, for example in military operations: platoon, linear wave front, V-formation, rectangular etc., each effective for particular predefined environments. For large size formations actual formation shapes will result from the concurrent effect of formation hold requirements and the particular environment constraints. Such self organized formations approach will not be based on predefined geometric shapes and is suitable for a mix of time varying number of autonomous vehicles. Autonomous vehicles considered in this project have local sensing and local actuating and, in most cases, are under-sensed and under-actuated.