The Research Of Control Allocation And Reconfiguration Of Multiple Control Surfaces Of Flying-wing UAVs

With the extensive research of new aerodynamic layout of UAVs being carried out and theincreasing complexity of the flight environment, control reconfiguration technique is becoming one ofthe main technical methods to guarantee the safety of UAVs. In this paper, a new fault detection anddiagnosis algorithm is designed, as well as a modified control allocation and reconfiguration method,according to the operating capability and fault characteristics of the aerodynamic redundant controlsurfaces of the sample fly-wing UAV. In this way, the control commands are allocated precisely andthe allocation can be reconstructed when the control surfaces fail.Firstly, we design a new fault detection and diagnosis algorithm based on multi-model parameterestimation, according to the typical failure types of the control surfaces of the sample UAV, to realizethe monitoring of the working state of the control surfaces by the on-line estimation of the dynamiccharacteristics parameters. And the failure type and extent can be given accurately and timely toprovide the basis of the reconfiguration.Secondly, according to the characteristics of the control surfaces, such as the position and rateconstraints, the large differences in operating effectiveness and the strong control coupling amongcontrol surfaces, the daisy chain method and the direct allocation is combined to solve these problems.This new approach divides the surfaces into several groups by considering the operating effectivenessfirstly and then a modified discrete time direct allocation is used for each group. It turns out that themodified method of control allocation can enhance the accuracy of the solution when constraints existand the control coupling is efficiently restrained.Then, a control re-allocation method is given based on the fault types and extents. The controlreconfiguration goal is obtained using the designed allocation algorithm by compensating the faultedcontrol surface and adjusting the operating effectiveness matrix.Finally, a real-time simulation environment is built by using the rapid prototyping technic,consisting of UAV model, control law, control allocation/reconfiguration and fault detection anddiagnosis module. The results show that the algorithms designed in this paper can provide the faultinformation of the UAV’s control surface correctly and timely to obtain the precise allocation of thecontrol commands.