SGER: Impact of Intermittent Networks on Team-based Coverage

态势感知越来越依赖于无人驾驶车辆团队通过持续的广域监视来提供信息。如何最好地协调多个自治代理的行动，无论是单独的还是团队的，以实现完整和及时覆盖的搜索目标，对于稳健的信息是至关重要的。不幸的是，机器人团队在覆盖任务中的应用有许多未解决的复杂性方面。具体地说，网络的传播限制限制了机器人团队使用数字网络促进合作的能力。在这个项目中，我们研究了在现实世界网络约束下的合作问题，重点是建模和实验。该研究项目旨在利用生物启发的合作范式：1)通过主动预测和管理基于角色分配的网络约束、基于邻近度的子团队形成和通过观察进行沟通来改进大规模搜索和监控；2)研究消息年龄和对性能的影响；3)研究消息重要性的贴近度。预计结果将与以前的工作一致，这些工作表明，减少沟通可以在某些领域和团队规模方面产生更好的合作。

Situational awareness is increasingly reliant on unmanned vehicle teams to provide information by constant wide area surveillance. Issues of how to best coordinate the actions of many autonomous agents, individually or in teams, in order to accomplish the search goal of complete and timely coverage are essential for the robust information. Unfortunately, the application of robot teams in coverage tasks has many unaddressed facets of complexity. Specifically, the propagation constraints of networks limit the ability for robot teams to facilitate cooperation using digital networks. In this project, we investigate the cooperation problem with a focus on modeling and experimentation in the presence of real world network constraints. This research project aims to use bio-inspired cooperation paradigms to: 1) improve large-scale search and surveillance by actively predicting and managing network constraints based on role assignment, proximity-based sub-team formation, and communication through observation, 2) study of message age and impact upon performance, and 3) study of proximity to message importance. It is expected that results will be in line with previous work that shows that less communication can result in better cooperation in certain areas and team sizes.