This paper studies the continuous-time constrained consensus of heterogeneous multi-agent networks with nonconvex input and velocity constraints, where each agent’s dynamic is modeled by a first-order or second-order integrator, the communication delays are assumed to be time-varying, nonuniform, bounded and the communication graph is changing over time. An improved distributed control algorithm with time-varying gains is proposed and the heterogeneous consensus is changed into the homogeneous consensus by a coordination transformation. Then, it is proved that the constrained consensus can be reached by analyzing six connection cases between agents if the joint graph has a directed spanning tree over each bounded interval. Finally, simulations are done to show the correctness of the theorem.

本文研究了具有非凸输入和速度约束的异构多智能体网络的连续时间约束共识，其中每个智能体的动态由一阶或二阶积分器建模，假设通信延迟随时间变化，不均匀、有界且通信图随时间变化。提出了一种改进的具有时变增益的分布式控制算法，通过协调变换将异构共识转变为同构共识。然后，证明如果联合图在每个有界区间上都有一个有向生成树，则可以通过分析代理之间的六个连接情况来达成约束共识。最后进行仿真以证明定理的正确性。