Secure state estimation and control for autonomous ships under cyberattacks

As the adoption of autonomous ships increases, their vulnerability to cyberattacks, particularly sensor attacks, has become a growing concern that threatens operational safety and efficiency. In response to this challenge, this paper presents a novel adaptive observer method designed to assess and quantify the severity of cyberattacks on autonomous ships. The proposed method operates in discrete-time and is specifically tailored to achieve simultaneous estimation of both the system's states and the size of the attack signals. The core of this approach is built upon two key principles: first, filtering the sensor measurements to isolate attack-related anomalies, and second, augmenting the measurement equation within the state-space representation to accommodate the effects of these attacks. This dual framework provides significant advantages over existing techniques, offering enhanced accuracy in estimating both system states and attack magnitudes. The methodology not only strengthens the resilience of autonomous ships against cyber threats but also paves the way for improved control strategies to mitigate the impact of such attacks, ultimately ensuring safer and more reliable autonomous maritime operations.