Integrated reconfigurable control and guidance based on evaluation of degraded performance

Abstract

The present M.Sc thesis is focused on studying an integrated reconfigurable control and guidance approach for recovering a small fixed-wing UAV from different actuator faults, which cover locked in place (stuck), floating and loss of effectiveness. Multiple simultaneous control surface faults can also be considered with this algorithm. The model of the UAV Aerosonde is used during the simulations, which are carried out in MATLAB/Simulink. First, a reconfigurable control system based on the control allocation technique is designed and tested for a variety of faults, specially locked in place control surfaces. Reconfiguration with this technique is shown to successfully recover the aircraft from the fault in some cases; while on the other hand, for other cases it is not enough to guarantee success on the planned mission. Therefore the system must, based on a measure of performance degradation, be able to detect whether the vehicle will be able to recover from the fault and carry out the full mission just by reconfiguring control or if at any point control deficiency is such that it will be necessary to change guidance command. For these situations where performance degradation is such that the prescribed trajectory cannot be achieved, a reconfigurable guidance system is developed, which is capable of adapting parameters such as the minimum turning radius and the look-ahead distance for obstacle avoidance to allow the vehicle to dynamically generate a path which guides the aircraft around the no-fly zones taking into account the post-fault reduced performance. Path following is performed by means of a non-linear lateral guidance law and a collision avoidance algorithm is implemented as well. Finally the integration of control reconfiguration and guidance adaptation is carried out to maximize probabilities of post-failure success in the mission. The link between both systems consists of using as a measure of performance degradation an error parameter from the control allocation algorithm. Although approximate, this results to be an efficient way of deciding the required degree of reconfiguration in the guidance command when an accurate prediction of the actual performance is not available. For the cases where it is, the system allows it as an input for more accurate reconfiguration.