Optimised sensor selection for control and fault tolerance: Comparison and some new results

Optimised sensor selection for control design is a non-trivial task to perform especially if the selection is done with respect to complex control requirements like reliability, optimised performance, robustness and fault tolerance. In this paper, a proposed framework is presented aiming to tackle the aforementioned problem. In this context, a Linear Quadratic Gaussian (LQG) controller is presented and applied to an Electro-Magnetic Suspension (EMS) system. Furthermore, the LQG solution is compared to a Multi-Objective (M.O.) H_∞ and H_∞ controller design via loop-shaping method using realistic simulations. A particular contribution is the use of Sensor Fault Accommodation Ratio (SFAR) in the LQG scheme providing useful conclusions on the optimised sensor selection for the EMS system. It is concluded that the framework can be extended to other industrial applications.