کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

RNN-Based Resilient Model Predictive Consensus Control for Nonlinear Multi-Agent Systems Under DoS Attacks: A Case Study in UAVs

نویسندگان :

Abdolah RoshanaeeDeh¹ Pariya Rezaei Chiyaneh² Iman Sharifi³

1- دانشگاه صنعتی امیرکبیر(پلی تکنیک تهران) 2- دانشگاه صنعتی امیرکبیر(پلی تکنیک تهران) 3- دانشگاه صنعتی امیرکبیر(پلی تکنیک تهران)

کلمات کلیدی :

Denial-of-Service،Multi-UAV،Resilient Control،Model Predictive Control

چکیده :

Abstract-This paper presents a RNN-based resilient model predictive consensus controller (RMPCC) for nonlinear multi-agent systems (MASs) operating over time-varying networks under denial-of-service (DoS) attacks, with a multi-UAV leader-follower case study. The closed loop switches between a nominal consensus law when links are healthy and an RNN-based RMPCC during DoS attack disruptions. A GRU trained offline approximates the finite-horizon RMPCC cost function, and a lightweight online search selects the control gain. We establish practical stability and exponential mean-square consensus under bounded DoS attacks and switching topology. Simulations on a Markovian switching-graph multi-UAV network confirm convergence and constraint satisfaction despite DoS attack disruptions.