非仿射多智能体系统的自适应神经网络控制

Adaptive neural network control of non-affine multi-agent systems

针对有向拓扑图下一类控制方向未知的非仿射非线性多智能体系统的输出一致性问题,综合运用中值定理、RBF神经网络及其特性、Nussbaum增益函数方法和动态面控制技巧,提出一种分布式自适应神经网络控制协议,保证跟随者的输出能与领导者的输出同步,跟踪误差能保持在零点的小邻域内。采用新的非线性滤波器代替传统动态面控制方法(CDSC)的一阶线性滤波器,改善控制性能。通过一致性分析及仿真例子验证了所提控制方法的有效性。

To solve the problem of output consensus control of leader-following nonlinear multi-agent systems under the directed communication topology,a distributed adaptive neural control protocol was proposed.Major design difficulties for this class of systems come from the non-affine system and the unknown control direction embedded in the unknown control gain function.A combination of mean-value theorem,Nussbaum gain function method and radial basis function(RBF)neural network approximation was employed to overcome these difficulties.To improve the control performance of conventional dynamical surface control(CDSC),a nonlinear filter was presented.The proposed control protocol guarantees that the outputs of followers can track that of the leader and the tracking errors can remain in a small neighbourhood of the origin.A simulation example was given to validate the effectiveness of the control methodology.