Prediction of the Flight Dynamics of Maneuvering Multirotor Aircraft

Abstract

A simulation environment is presented that predicts the flight trajectory of a maneuvering multirotor aircraft using a purely physics‐based approach without the need for a priori flight test data. The flight dynamics model determines the motion of the aircraft based on the total loads and commanded motor speeds. The aerodynamic loads of the rotors are predicted using a modified blade element momentum theory (BEMT)–based approach that considers nonuniform inflow conditions at the rotor discs. In addition, the aerodynamic loads of the remaining aircraft components are estimated using a load decomposition. Flight test data of an AscTec Pelican quadcopter were used to evaluate the prediction quality by comparing it with the vehicle tracks recorded in flight tests. As the flight changes from hover, the present approach shows significant prediction improvements over a simple KΩ2 approach. Specifically, when comparing the number of successful prediction timesteps into the future, the BEMT‐based approach showed, on average, 44.4% longer successful predicting for positional velocities and 85.3% longer for predicting body rates. In addition to its numerical accuracy, the simulation environment is computationally efficient and thus ideal for design studies of flight controllers. The codes associated with the simulation environment are open source.