基于事件触发的复杂网络的同步、一致性和蜂拥控制的研究

With the deepening of research work on the theories and applications of complex networks, how to design appropriative mutual communicating protocols among the nodes has attracted a lot of researchers' concerns. It is required in the prior literatures of complex networks which are sampled periodly that all nodes should send the sampled data packets at the same time. Obviously, it will lead to longer time delays, a greater rate of data packet dropout, or even cause the network congestion, since the current network bandwidths are limited. In this proposal, we present an event-triggered mechanism to minimize the number of packets transmitting through the network, such that the configuration of the limited resources would be optimized and the reliability of the complex networks could be improved. We would focus on the topic that analysis and synthesis of even-triggered complex networks. Our research will begin from the easy to the difficult and complicated. Firstly, the consensus and synchronization of event-triggered complex networks involving simple time delays in the communication channels will be studied.The basic event-triggered mechanism will be designed. Secondly, we will model the event-triggered complex networks transmitting signal through the general communication channels, in which the time delays, packet dropout and stochastic disturbances will be considered. The consensus problem and synchronization control will be considered. Then, the theory framework of researching on event-triggered complex networks will be further developed. Thirdly, the performance evaluation schemes for the event-triggered complex networks will be proposed, and then, the flocking control, adaptive cooperative control and some control optimization methods for the complex system will be investigated, respectively. At last, the cross-layer optimization design method will be studied between the communication protocol parameters and the ones in the event-triggered function. The results achieved in this proposal will provide new ideas and theory supports for some artificial complex networks, such as the alignment of unmanned aerial vehicles, the intelligent driving systems, and so on.

随着复杂网络理论与应用研究的深入，节点间的通信问题受到人们的广泛关注。先前基于周期发送数据的复杂网络要求所有节点同时发送采样数据。由于网络带宽的限制，这势必会造成更长的通信时滞、更大的丢包率，甚至造成网络拥塞。项目拟开展基于事件触发的复杂网络的研究工作，旨在尽可能减少网络中数据包的传输数量，优化有限的资源配置，提高系统的可靠性。研究内容由浅入深，以简单通信环境下事件触发复杂网络的一致性分析和同步控制为起点，探讨事件触发机制的设计问题；继而研究常规通信网络下事件触发复杂网络的建模、一致性分析和同步控制问题，构建事件触发复杂网络的理论研究框架；在建立系统性能评价指标的基础上，深入探讨系统的蜂拥控制、自适应协同控制的设计及相关控制优化问题；最后研究网络通信协议参数和事件触发函数的跨层优化设计问题。项目的研究成果将为无人驾驶飞行器编队、智能汽车驾驶系统等人工复杂网络的设计提供新的思路和理论依据。

复杂网络的研究近年来受到广泛关注。人们尝试通过不同形式的网络对大量由自然的、社会的或者人工的元素组成的复杂系统加以刻画，寻求客观实在呈现表象的内在机理和模式，发现影响和支配系统性态和演化规律的本质原因，并探索干预或改造网络动力学行为的方法。另一方面，如何节省有限的网络资源也是研究者关注的热点。课题围绕基于事件触发的复杂系统的分析与综合问题展开，项目的研究计划要点均已完成。课题组研究了简单网络通信情况下事件触发复杂网络的相关理论；分析了一般网络通信环境下事件触发复杂系统的动态及其相关控制问题；建立基于事件触发的复杂网络协同控制性能的评价机制；研究了基于事件触发复杂网络的优化问题；并搭建实验平台，验证所提模型和控制策略。在分析方法上，项目也取得了一些研究成果，针对系统中非线性约束，推广了现有奇异值分解引理，并应用到控制器设计中，减小了系统的保守性；针对网络信号传输会改变正系统正性，将自由变量引入系统建模过程，首次给出了基于事件触发的网络化正系统的控制设计方法，该论文为网络化正系统研究方面的第一篇论文。另外，在项目研究过程中我们也对相关领域进行了探讨，例如研究了挠性航天器的鲁棒控制、自适应控制等问题。项目的研究成果为近年来网络化复杂系统研究中涌现出的新问题提供参考，例如无线信道的Rayleigh衰减对系统群集行为的影响等。项目研究成果为无人驾驶飞行器编队、智能交通系统等人工复杂网络的设计与优化提供新的思路和理论依据。标注项目资助的研究成果，发表期刊论文29篇，其中SCIE收录13篇，SCIE源刊接受论文7篇，发表会议论文17篇。申请发明专利3项。在此资金资助下，项目负责人及其项目组成员参加国内外的学术会议，进行学术交流，力求掌握学科前沿最新进展和动态。在注重交流的同时，也积极发展本研究小组的学术梯队，现培养博士研究生3人，在读4人，硕士生研究生8人，在读4人。