Event-triggered data-driven secure control for unmanned marine vehicles under hybrid attacks

Abstract

Traditional trajectory tracking control methods rely excessively on precise models of unmanned marine vehicles (UMVs). However, due to complex marine environmental disturbances, it is challenging to obtain accurate models for UMVs. Therefore, a data-driven approach is employed for UMVs trajectory tracking control. An improved prediction compensation method is proposed, which incorporates an arctangent function for pseudo Jacobian matrix updating to achieve more accurate predictions compared to traditional methods. Moreover, an event-triggered (ET) mechanism incorporating a dead-zone operator is designed. Compared to fixed triggering conditions, this mechanism balances system performance and triggering frequency by adjusting the parameter of the dead-zone operator, thereby enhancing triggering flexibility. In addition, an improved extended state observer (ESO) is designed by integrating the successfully transmitted output and the predicted output. It effectively addresses the impact of hybrid attacks and ET on the ESO in the disturbed system, enabling the estimation of lumped disturbances. This paper introduces a novel secure control strategy that integrates an attack compensation mechanism with an improved ESO, ensuring that UMVs can reliably track the desired trajectory despite external disturbances and hybrid attacks. The entire design process relies on input/output data only, and simulation comparisons validate the effectiveness of the proposed methods.