Actuator multiplicative and additive simultaneous faults estimation using a qLPV proportional integral unknown input observer

This paper introduces a technique for simultaneous estimation of additive and multiplicative faults in the actuators of nonlinear systems represented by quasi-linear parameter varying (qLPV) models based on a proportional-integral unknown input observer. The qLPV model, structured with a tensor product, allows for optimized flexibility of the observer gain. A distinguishing aspect of our method is the novel approach to nonlinearity, which is not only recast as a convex sum but also in the input vector. The study comprehensively analyses the robustness and convergence conditions through Lyapunov stability evaluation. A robust $H_{\infty }$ performance criterion is incorporated to minimize the influence of measurement noise and disturbances. As a result, a set of linear matrix inequalities are obtained. Two examples are examined to demonstrate the practical applicability and efficacy of the proposed method, highlighting the observer's performance under the actuator faults.