Distributionally Robust Optimization Adaptive Event‐Triggered SMPC for DC‐DC Converters Subject to Unknown Disturbances and DoS Attacks

ABSTRACT In this paper, we present adaptive event‐triggered distributionally robust optimization stochastic model predictive control (AET‐DROSMPC) applied to DC‐DC converters subject to unknown disturbances and denial of service (DoS) attacks. The DoS attacks, causing communication interruptions on the controller and actuator (C‐A) channel, is described by using Bernoulli variables. While stochastic model predictive control (SMPC) has been extensively studied, existing approaches mostly focus on periodic stochastic model predictive control (PSMPC) and self‐triggering stochastic model predictive control (SSMPC) for systems with bounded disturbances or subject to Gaussian distribution. To address chance constraints and disturbances more effectively, we introduce the distributionally robust optimization (DRO) to transform the optimization problem with chance constraints into a convex optimization problem using second‐order cone (SOC) equivalence. Moreover, adaptive event‐triggered mechanism (AETM) is devised to reduce unnecessary sampling, lower updating frequencies of control input, and ultimately decrease the computational burden of the system, addressing the lack of consideration for exact sampling times and system computing burden. The study rigorously establishes the recursive feasibility and stability of the optimization problem subject to DoS attacks. Finally, a application is conducted to demonstrate the effectiveness and advancements of the proposed algorithm.