Optimal attitude and flight vector recovery for large transport aircraft using sequential quadratic programming

Loss of control prevention for aircraft remains an active area of research to improve aviation safety. Two separate numerical methods are presented here and were used to generate attitude and flight vector recovery solutions for a simulated aircraft model in an initial upset condition. The problem is posed as an optimal control problem and both methods' general approach uses a reduced-order mathematical model of a sub-scale jet airliner simulation model. The first method discussed uses a method called dynamic programming (DP) to generate a discrete closed-loop solution and the second method uses direct transcription and sequential quadratic programming (SQP) to generate an open-loop solution. The two methods' trajectory solutions are compared with an illustrative example and it was found that both methods' solutions use the same maneuver strategy to recover the aircraft model's trajectory back to wings level flight. In conclusion the DP is limited to a reduced model to remain tractable while the SQP can use a more accurate and representative model, and the DP method is computationally expensive offline while the SQP method is expensive online.