Stochastic Model Predictive Control for Airspeed Optimization Using Successive Convexification

摘要

The present paper addresses the problem of airspeed profile optimization along a pre-assigned trajectory and newly proposes a real-time method for robust airspeed planning under uncertainties of wind forecasts and the response time of the lower-level closed-loop system of the airspeed control. In order to reduce the number of decision variables of the airspeed profile optimization, the method proposed herein adopts the stochastic model predictive control (MPC) based on the unscented transform (IT) for uncertainty modeling. In addition, to make each step of the MPC tractable, we apply successive convexification of the non-convex constraints incurred by the IT via local linearizations and relaxations to convex quadratic inequalities. Through numerical simulations, the effectiveness of the proposed approach is confirmed in terms of the fuel efficiency, punctuality of the arrival, low pilot workload, and high computational speed.