Data-Based Design of Robust Fault Isolation Residuals Using LASSO optimization

In this paper a data-based approach is proposed for the design of structured residual subsets for the robust isolation of sensor faults. Linear regression models are employed to estimate the faulty signals and to build a set of primary residuals. A L1-regularized least square parameter estimation technique is used to identify the model parameters and to enforce sparsity of the solutions by increasing the regularization weight. In this way it is possible to generate a large set of residuals generators having different fault sensitivity. Then, a residual selection procedure, based on fault sensitivity maximization, is proposed to select a minimum size subset of residuals that allows the isolation of the faulty sensor subet. The proposed method has been validated designing a fault isolation scheme for 6 aircraft sensors using multi-flight experimental data of a P92 Tecnam aircraft.