Model-based approaches for the airspeed estimation and fault monitoring of an Unmanned Aerial Vehicle

This paper presents a methodology for the design of a diagnostic system for the detection of faults occurring on the airspeed sensor of a remotely controlled semi-scale YF-22 monitoring aircraft developed at West Virginia University (WVU). Based on the mathematical model of the UAV, the dynamic equation of the angle of attach is used for the estimation of the value of the airspeed. Two methodologies are here proposed. The first approach (model-based method) assumes a perfect knowledge of the aircraft; in other words, all the geometric and inertial parameters as well as the aerodynamic coefficients are assumed to be known and correct. In the second approach the parameters governing the equation are instead identified via LS optimization from a batch of simulated flight data. Assuming the availability of the measurement of the whole state vector except the airspeed (V), the angle of attack equation results quadratic in V and thus can be easily solved on-line to provide a real-time estimation of the airspeed. The effectiveness of he proposed estimators was validated within a sensor failure detection scheme where the airspeed estimation is employed to generate the diagnostic signal for detecting possible additive faults on the airspeed sensor. A `cusum' detector was employed to detect the occurrence of the failure. The performance of the proposed methodology have been evaluated through a simulation study carried out on a nonlinear dynamical model of the WVU YF-22 aircraft.