Control and Optimization for Event-triggered Networked Autonomous Multi-agent Systems

众所周知的短语-冠军团队将击败冠军团队-提供了关于为什么多智能体系统的合作控制研究被视为当前和未来工程技术中最有前途和最令人兴奋的方向之一的见解。机器人代理团队相互合作和互动，以完成搜索和救援，大面积监视或环境监测等任务。其关键是为每个代理开发分布式控制算法，共同实现全局目标。本项目将推进分布式控制算法的现有成果，并通过研究实现算法的能量成本做出新的贡献。首先，基于事件的控制框架将用于通过仅在需要时感测和致动来减少能量消耗，而不是大多数现有的算法，其连续地致动和感测。 此外，与现有的结果不同，我们将研究具有复杂机器人动力学的代理。其次，该项目将考虑在实现合作任务中消耗的能量。不同的协调任务，如编队控制，最优控制器的设计，以减少能源消耗，目前只有共识任务已被考虑。除此之外，还将纳入能源成本相关代理技术的复杂措施。由PI Yu和PI Hirche领导的团队在技能上的互补性意味着这两个团队非常适合在这个项目上合作。PI Yu和HDU团队在编队控制和分布式协调方面具有特殊优势，PI Hirche和TUM团队以其对最优控制，基于事件的控制和机器人技术的理论贡献而闻名。鉴于过去的成功合作，以及世界领先技能的互补，我们相信这个项目也将成功地产生巨大的研究贡献。

The well-known phrase - a champion team will defeat a team of champions - gives insight as to why research on cooperative control of multi-agent systems is seen as one of the most promising and exciting directions in current and future engineering technology. Teams of robotic agents cooperate and interact with each other to complete tasks such as search and rescue, surveillance coverage over a large area or environmental monitoring. Key to this is development of distributed control algorithms for each agent that collectively achieves the global objective.This project will advance existing results in distributed control algorithms and make new contributions by studying the energy cost of implementing the algorithms. Firstly, the framework of event-based control will be used to reduce energy consumption by sensing and actuating only when required, as opposed to most existing algorithms, which actuate and sense continuously. Furthermore, unlike existing results, we will study agents with complex robot dynamics. Secondly, the project will consider the energy consumed in achieving the cooperative task. Optimal controllers for different coordination tasks, such as formation control, will be designed to reduce energy consumption; currently only consensus tasks have been considered. Going beyond this, sophisticated measures of energy costs related agent technologies will be incorporated.The complementarity in skills of the groups led by PI Yu and PI Hirche means the groups are perfectly suited for collaboration on this project. PI Yu and the HDU team have particular strengths in formation control and distributed coordination and PI Hirche and the TUM team are known for their theoretical contributions to optimal control, event-based control and robotics. Given past successful collaborations, and the complementary of world-leading skills, we believe this project will also be successful in producing great research contributions.