EAGER: Virtual Motion Camouflage based Subspace Optimal Control for Real-Time Trajectory Planning

这项探索性研究（EAGER）奖的研究目标是调查、更好地理解和验证一种新颖独特的子空间最优控制方法，该方法将为推进单个或协作系统实时轨迹规划关键领域的知识状态奠定基础。本文将研究一种基于仿生运动伪装现象和高阶离散化方案的新方法，从而大大降低求解约束非线性最优轨迹问题的计算量。本文将研究一种系统的工具，用于在获得解后判断解的最优性，并遵循一个理论证明的指导方针，即如何事先选择虚拟猎物运动，以使一个？子空间可以先验地构造。在这个EAGER项目中，将使用两个示例问题来验证方法：协作电子战斗飞行器（ECAVs）问题和移动机器人碰撞轨迹规划问题。作为一种潜在的变革性技术，所提出的研究不仅为实时轨迹规划提供了一种创新方法，而且还有助于以标准约束非线性最优控制形式建模的广泛其他应用。更重要的是，本研究结果将促进自然界中存在的生物运动策略在实际工程或科学问题中的应用。本研究有望对理论和应用产生影响，并弥补两者之间的差距。项目负责人和研究生将参加由NSF CMMI组织的论坛，在整个项目期间传播我们的创新和发现。研究成果将通过项目报告、期刊出版物和同行反馈进行评估。生物现象和研究子任务和成果将用于丰富本科动力学课程和研究生最优控制课程。新的教材将被链接到国家科学数字图书馆。

The research objective of this EArly-concept Grants for Exploratory Research (EAGER) award is to investigate, better understand, and validate a novel and unique subspace optimal control methodology that will potentially build a foundation for advancing the state of knowledge in the critical area of real-time trajectory planning for single or cooperative systems. A new method, based on the bio-inspired motion camouflage phenomenon and higher order discretization schemes, will be investigated such that the computational cost in solving constrained nonlinear optimal trajectory problems can be dramatically reduced. A systematic tool of judging the optimality of the solution after it is obtained will be investigated and followed by a theoretically proven guideline on how to select the virtual prey motion beforehand such that a ?good? subspace can be constructed a priori. In this EAGER project, two example problems will be used to validate the methodologies: a Cooperative Electronic Combat Air Vehicles (ECAVs) problem and a mobile robot collision trajectory planning problem.As a potentially transformative technique, the proposed research will not only provide an innovative approach for real-time trajectory planning but also contribute to a wide range of other applications that can be modeled in a standard constrained nonlinear optimal control form. Even more important, the results of this research will promote the implementation of biological motion strategies existing in nature to practical engineering or scientific problems. This research is expected to impact both theories and applications and bridge the gap between them. The PI and graduate students will attend forums organized by the NSF CMMI to disseminate our innovations and findings throughout the project period. Research findings will be evaluated by project reports, journal publications, and peer feedback. The biological phenomenon and research subtasks and findings will be used to enrich the undergraduate dynamics course and the graduate optimal control course. New teaching materials will be linked to the National Science Digital Library.