Neural Network Observer Based Consensus Control of Unknown Nonlinear Multi-agent Systems with Prescribed Performance and Input Quantization

Shi Zhengqing,Zhou Chuan,Guo Jian
DOI: https://doi.org/10.1007/s12555-020-0326-8
2021-01-01
Abstract:This paper investigates the consensus tracking problem with predefined performances requirements for a class of unknown nonlinear multi-agent systems with hysteresis quantizer and external disturbances under a directed graph topology. Neural network observers are designed to estimate unmeasurable states and the the consensus tracking problem with performance requirements is transformed to a stabilization problem by prescribed performance error transformation schemes. The novel consensus protocol can be applied to a more general class of nonlinear multi-agent systems since the Lipschitz condition is avoided and state information is not required. It is strictly proved that all signals in the closed-loop systems are cooperatively uniformly ultimately bounded and both the transient and steady performances of the consensus tracking satisfy prescribed performance requirements. Finally, two numerical examples are presented to validate the effectiveness of the proposed strategy.
What problem does this paper attempt to address?