Neural-Network Bilinear Gain-Scheduling Control for B747-100 Longitudinal Dynamics

Ezzeddin M. Elarbi, Dina Shona Laila, Nadjim Horri

Research output: Contribution to conferencePaper


A Gaussian radial basis function neural network was applied to control the longitudinal dynamics of Boeing 747-100. Bilinear interpolation calibrates the responses to allow for gain scheduled control within the whole flight envelope. The baseline responses are produced from the simulations based on Linear Quadratic Regulator (LQR) control combining elevator and throttle commands at three trim conditions on Mach number (M) and altitudes: (M=0.2,0m), (M=0.5, 6096m) and (M=0.9, 12192m). The Latin hypercube is employed to uniformly discretize the flight envelope with Mach numbers ranging from 0.2 to 0.9 and altitudes from 0 to 12192m. The steady state response surfaces of longitudinal references are achieved based on gains smoothly satisfying closed loop flying qualities. The elevator and throttle inputs can be adjusted to maintain the center of gravity within admissible limits over the entire flight envelope. The pitch trim condition can be autotuned by autothrottle for a given center of gravity location.
Original languageEnglish
Publication statusSubmitted - 25 Jan 2019

Fingerprint Dive into the research topics of 'Neural-Network Bilinear Gain-Scheduling Control for B747-100 Longitudinal Dynamics'. Together they form a unique fingerprint.

Cite this