网络环境下基于分布式事件触发采样机制的非线性多智能体系统一致性控制研究

Event-triggering control based strategy can effectively reduce the traffic load of communication network, which is very suitable for solving the design problem of consensus protocol of constrained multi-agent systems with limited energy and bandwidth in communication process. Within a more general network topology structure, this project investigates the problem of consensus control and application of nonlinear multi-agent systems with a distributed event-triggered sampling scheme under the network environment including data quantization, data encoding, data decoding, transmission delay of network, packet loss and controller input saturation simultaneously by introducing a suitable nonlinear function to model the inherent dynamical behavior of agents and constructing a generally applicable mathematical model. Through the design of suitable distributed event-triggered functions and excitation conditions to determine the individually independent event-triggered time series and by conceiving appropriate control strategies only based on the individual sampled data at their own triggering time instants in each agent's neighborhood, some sufficient consensus conditions of multi-agent systems are obtained and the underlying explicit relationship between the parameters characterizing the network environment and the parameters of networked control system are also found. The completion of this project will enrich and develop the theory of high performance consensus protocol design of nonlinear multi-agent systems with limited capacity of the communication channel and limited computing power of agent’s processor, and will effectively overcome that the traditional continuous state control strategy is difficult to apply directly to the actual networked systems.

基于事件触发控制策略能减轻网络通信负担，适用于解决通讯过程中能量有限且通信带宽受限的多智能体系统一致性协议设计问题。本项目考虑在更一般的有向网络拓扑结构下，引入能建模智能体自身固有动力学行为的非线性函数，建立具有普适意义的数学模型，研究在数据量化、编码、解码，网络传输时延，数据丢包和控制器输入饱和同时存在的网络环境下基于分布式事件触发采样的非线性多智能体系统一致性控制及其应用问题。通过设计合理的分布式事件触发函数与激励条件，确定智能体彼此独立的事件触发时间序列，利用邻域内其它个体在各自事件触发时刻点处信息，设计相应的控制策略，获得多智能体系统达到一致性的充分条件，寻找表征网络环境相关参数与网络化控制系统参数之间的关系。本项目的完成将丰富和发展在通信信道容量受限和智能体处理器计算能力有限下，非线性多智能体系统高性能一致性协议设计问题，克服传统连续状态控制难以直接应用到实际网络化系统的不足。

基于事件触发控制策略可以减轻网络通信负担，适用于解决多智能体系统一致性协议通讯过程中能量有限且通信带宽受限的问题。本项目主要研究内容侧重于考虑在更一般的有向网络拓扑结构下，引入能建模智能体自身固有动力学行为的非线性函数，建立具有普适意义的数学模型，研究在复杂网络环境下基于分布式事件触发采样控制的非线性多智能体系统的一致性控制及其应用问题。通过项目组成员之间的团结合作，本项目建立了更加完善的基于分布式事件触发采样控制的非线性多智能体系统一致性数学模型，给出了高性能一致性控制协议，事件触发函数和激励函数设计方案，并进而提出了一致性分析和不存在芝诺行为的全新理论框架；在数据量化、数据编码、数据解码、网络传输时延数据丢包、控制器输入饱和等因素同时存在的复杂网络环境下，提出了相关网络参量的设计方案和表征网络相关参数的最大理论上界与一致性精度之间的解析关系。本项目将丰富和发展在通信信道容量受限和智能体处理器计算能力有限下，非线性多智能体系统高性能一致性协议设计问题，克服传统连续状态控制难以直接应用到实际网络化系统的不足。本项目为探讨基于事件触发多智能体系统一致性的研究奠定了坚实的工作基础。同时对于认识复杂网络环境对多智能体系统一致性的影响提供了重要的启示性线索。