Nonlinear Control Design And Control Allocation For Tailless Flying Wing Aircraft

In this work we study several issues related to the nonlinear control of an overactuated flying wing aircraft.There are three major topics which are studied here.Robust nonlinear dynamic inversion(NDI),robust nonlinear control allocation,and control interaction or control coupling.The directional stability and control is crucial for the low-speed flight of a flying-wing aircraft.The split drag-rudders are well known devices used to provide directional stability and control in a flying-wing aircraft.As opposed to conventional rudders,the control efficiency of split drag-rudders is typically low for small deflection-angles and the influence on yawing moment is nonlinear.Such characteristics limit the control capability of split drag-rudders at low speed flight with large angle of attack.In this work,a new and simple method is presented to improve the control efficiency of split drag-rudders at low speed flight with large angle of attack.The method results in a strictly differential configuration of split drag rudders that operates around a certain bias or offset input.For last three decades,the NDI has become increasingly popular technique for the flight control design.This is due to the fact that an NDI controller provides good performance,over the complete flight envelope,with ideally requiring no gain-scheduling.Modern high angle of attack fighter aircrafts F-18,F-35,and crew return vehicle X-38 are the examples of high performance aircrafts,in which NDI was successfully used.However,it is difficult to find any significant evidence of use of NDI for the control of a flying wing aircraft.Flying wings are usually designed for high altitudes and long ranges.They are also required to operate at wide range of speeds.This wide variation in altitude and speed significantly alters the dynamics of aircraft in different operating regimes.Thus,it is very plausible to use NDI for the control of a flying wing aircraft.In this work a new Robust NDI design is presented for the flying wing aircraft.The complete NDI control law consists of three parts.The onboard aircraft model(OBAC)or simply the internal model,dynamic inversion control law,and control allocation method.Traditionally,the internal model is implemented in the form of polynomial functions,which are obtained through least square polynomial fitting.However,the aerodynamic data,which is obtained through wind tunnel testing or computational fluid dynamics(CFD)simulation,is inherently piecewise linear function of flight states and control surface deflections.The canonical piecewise linear representation is a popular technique to model piecewise linear functions.This technique has been used extensively for the analysis and simulation of electrical circuits.In this work,we exploit the piecewise linear nature of aerodynamic data and propose a new method of constructing the internal model,which is based on canonical piecewise linear representation.Uncertainty always exists in the internal model,even in the most accurately approximated one.Since,the standard NDI control relies on the accurate internal model,its performance may be degraded if uncertainty is present in the model.This problem has been addressed previously by combining robust techniques like H_?synthesis and?-synthesis with the NDI.A rather different approach,called incremental NDI(INDI),uses the angular acceleration feedback to make the NDI control law insensitive or robust to the parametric uncertainties.The INDI techniques proposed in past were designed by assuming that the control effectiveness is the linear function of control surface deflection.Furthermore,the previous INDI designs were proposed for the conventional aircrafts with three control variables aileron,elevator and rudder.In this thesis,we extend the existing INDI control design to the case of an overactuated aircraft with nonlinear moment versus deflection relationships.This is done by integrating a nonlinear control allocation algorithm,like redistributed pseudoinverse(RPI)with locally affine effectiveness,with the main INDI control law.Consequently,the nonlinear control allocation algorithm integrated with the main INDI control law results in the robust nonlinear control allocation.In landing phase,usually a tight flight path control is required.In this phase it is reasonable to track the flight path angle instead of pitch angle.In this thesis,a new longitudinal control law is also presented,in which the inner loop and control allocation is based on INDI control;however,the outer longitudinal control loop uses the backstepping based flight path angle control.Furthermore,the effects of sideslip angle on the effectiveness of the split drag rudders are modeled within the nonlinear control allocation algorithm,and the control allocation performance is tested in simulation for landing phase in presence of lateral wind disturbance.Finally,a solution to the problem of nonlinear control allocation in presence of control coupling is presented.Traditionally,the control coupling is ignored in the control allocation problem.However,there are the cases in which it may not be ignored.The only available technique for incorporating control coupling effects in the control allocation algorithm is based on the second-order Maclaurin Expansion,which can approximate the coupling only near the origin.In this thesis,we present a new modeling technique,which is based on the local approximation of non-separable function using two-dimensional first-order Taylor Expansion.The technique makes it possible to incorporate,in the control allocation algorithm,the control coupling effects for the complete range of control surface deflection,thus resulting in accurate nonlinear control allocation.