Compensating Latent Nonlinear Dynamics for Practical Consensus Control

In this paper, we propose a new kernel-based method for compensating latent nonlinear dynamics for consensus control in multiagent systems. Although kernel regression is a well-known and thoroughly studied technique, recent research has shown its significant non-asymptotic potential. Under general conditions, we show the convergence of the proposed approach by stability analysis and show that applying kernel regression compensation for consensus control leads to synchronization of the agents within high probability error bounds.