Comparative study of fault estimation for 3-DOF helicopter

Fault estimation plays an important role in health monitoring as well as fault tolerant control, which can help ensure the reliability and safety of helicopter system. As the helicopter is a typical nonlinear, strong coupling and unstable system, its fault estimation design is challenging. Besides, it is also of great risk and cost to conduct experimental validation of different fault estimation schemes for real helicopter system. Thus, this paper is intended to conduct comparative research of fault estimation for helicopter system with table-mount experimental platform. Based on helicopter's dynamic model, four observer-based fault estimation approaches are presented, in which Luenberger observer, Kalman filter, unknown input observer, and modified proportional-integral-derivative observer design are adopted separately. For the modified proportional-integral-derivative observer based fault estimation design, measurement noise is specifically considered. Based on Quanser's 3-DOF helicopter system, both simulation and experimental tests are carried out to show the performance differences of these fault estimation approaches.