【摘要】 针对一阶和二阶智能体组成的异构多智能体系统，研究了固定和切换拓扑下异构系统分组一致性问题。在设计分组一致性协议中加入了特殊变量并设计各组智能体的收敛位置，让各组智能体分别向着各自目标点聚集，各组内部智能体状态达到一致。方法上主要将异构系统状态矩阵进行变换，将异构系统的状态矩阵转化成同阶系统矩阵，让异构系统的分组一致性问题等价于同构系统的一致性问题，之后分析出固定拓扑下系统实现一致性所需条件。当系统含有切换拓扑时，将系统状态矩阵拆分成有限个非周期不可约分矩阵的乘积，通过证明有限个矩阵乘积为常数向量，即系统矩阵状态值最终收敛。在切换拓扑图的并集包含一个有向生成树下对系统矩阵进行分析得到系统实现分组一致性的条件。最后数值仿真验证了理论的正确性。更多还原

【Abstract】 For heterogeneous multiple smart-agent systems composed of first and second order agents,the grouping consensus of heterogeneous systems in fixed and switching topologies is studied. Special variables are added into the design of grouping consensus protocol and the convergence positions of agents are designed to make agents converge towards their target points and achieve the consensus of agent states within each group. In the method,the state matrix of the heterogeneous system is transformed into the same order system matrix,so that the group consensus of the heterogeneous system is equivalent to the consensus of the homogeneous system. After that,the required conditions for the system to achieve consensus under the fixed topology are analyzed. When the system has a switching topology,the state matrix of the system is divided into the product of a finite number of nonperiodic and nondivisible matrices,and the product of a finite number of matrices is proved to be a constant vector,that is,the state value of the system matrix converges finally. The condition of grouping consensus is obtained by analyzing the system matrix under the union of switched topology containing a directed spanning tree. Numerical simulation verifies the correctness of the theory.更多还原