Incremental Backstepping Control for Helicopters

Abstract

Due to their highly coupled and non-linear behavior, helicopters seem excellent subjects to apply non-linear control theory to. Furthermore, the rotor is operating in its own wake, leading to complex aerodynamics. A Command Filtered Incremental Backstepping controller has been applied to a simulation model of an MBB Bo-105 hingeless rotorcraft. The model incorporates the Pitt-Peters inflow model to calculate the inflow variations of the main rotor. Incremental controllers rely on sensor measurements of state derivatives instead of model knowledge, making them robust to modelling errors. However, some states of the helicopter model, such as blade flap angles and rotor inflow, cannot be measured in real life. Therefore a process called residualization and synchronization is used to remove these states from the controller model and compensate for their dynamics in a synchronization filter. This process has already been performed for the flap angles. In this thesis the inflow states are also residualized and added to the synchronization filter. Furthermore, the Pitt-Peters inflow model of the simulation model is updated with the Keller correction to better simulate off-axis response to control input. Modelling the off-axis response of helicopters is notoriously difficult and is often of the wrong sign compared to experimental data. Having a more precise helicopter model is key to perform piloted simulation or research to non-linear controllers. Although the updated inflow model did alter the inflow states, changes in the helicopter dynamics remained very limited. Furthermore, the application of residualization was successful but synchronizing the rotor inflow did not improve controller performance.