Adaptive Event-Triggered Consensus for Unknown Nonlinear Multi-Agent Systems with Limited Bandwidth

This paper proposes an adaptive event-triggered consensus control framework for multi-agent systems (MASs) with unknown nonlinear dynamics and limited communication bandwidth. First, an observer-based radial basis function neural networks (RBFNNs) compensator is designed to approximate unknown nonlinearities through online weight adaptation. Second, a dynamic encoding-decoding protocol (EDP) is introduced to compress transmitted data between different agents, avoiding quantizer saturation while preserving privacy. Third, an adaptive event-triggered controller is developed, employing dynamic threshold conditions to reduce triggering frequency compared to conventional time-triggered schemes. Theoretical analysis proves uniformly ultimately bounded (UUB) consensus, rigorously excludes Zeno behavior, and guarantees boundedness of transmitted data. The work provides a systematic framework for distributed control in bandwidth-limited, uncertain environments.