State Observation and Unknown Input and Output Disturbance Estimation in Networked Control Systems

In modern control systems, actuators and sensors are connected over communication networks. Such systems, referred to as networked control systems, can be subjected to various disturbances. These disturbances can have form of malicious attacks on the communication channels between the plant sensors and the controller and between the controller and the actuators. To protect the networked control system against such attacks, detectors of incoming attacks are needed. Attacks on the plant actuators are modeled as unknown bounded inputs, while attacks on the plant sensors are represented as output disturbances.

Continuous-time (CT) and discrete-time (DT) unknown input observer (UIO) architectures are developed to estimate the state, unknown inputs, and output disturbances in networked control systems. Adaptive CT schemes for unknown input and state estimation are proposed. Novel DT state and unknown input observers are proposed for a class of nonlinear networked control systems whose nonlinearities can be characterized by incremental multiplier matrices. Then, DT unknown input and output disturbance estimators are developed for the detection of attacks on the plant input and output channels. Delayed unknown input and output disturbance estimators are proposed for DT networked control systems for which the matrix rank condition for the existence of UIOs is not satisfied. An observer-based decentralized control design method is proposed for networked control systems where the communication network is modeled as a pure time-delay.

The results obtained can be applied to the observer-based decentralized control of networked control systems in the presence of time-delays and disturbances resulting from the presence of the communication networks.