Abstract
Abstract
In this study, a novel intelligent aircraft head model (AHM) system integrated with a piezoelectric cantilever array (PCA), and a convolutional neural network (CNN) algorithm for precisely predicting the flight parameters, which include the airspeed (
v
air
), angle of attack (
ψ
), and angle of sideslip (
γ
), is proposed. A cone-shaped frame with the novel ring PCA surface configuration is designed for adapt to the aerodynamic requirements of most flight vehicle surfaces. Simultaneously, a piezoelectric bimorph cantilever (PBC) vibration mathematical model is established by the electromechanical coupling equation via numerical solution of MATLAB. The analytical results demonstrate the PBC’s vibration mechanism with the cone-shaped frame. Several comprehensive experiments are performed to obtain the sensing characteristics of flight parameters and to verified the above CNN algorithm with AHM-PCA prototype. The analytically predicted values match well with the experimental results in terms of the response voltage (
U
(
t
)) and vibration amplitude (
y
(
t
)) with respect to
v
air
.
Meanwhile, the experimental results indicate that the
ψ
and
γ
have linear sensing characteristics in the range of -20° to 20°. Furthermore, comparisons of analytical and experimental results show that the absolute errors of the
v
air
,
ψ
,
and
γ
are less than 0.13 m/s, 0.15°, 0.12°, respectively, utilizing the CNN algorithm. Therefore, the superiority of the AHM-PCA's performance on flight parameters is demonstrated, and the accuracy of
v
air
,
ψ
,
and
γ
exceeds the previous state-of-the-art (SOTA) method, and the advantages reach 208% (0.13 m/s vs 0.27 m/s), 580% (0.15° vs 0.87°), 225% (0.12° vs 0.27°), respectively. The flight parameters predicted with the piezoelectric smart configuration and CNN algorithm prove potential application in the flight parameter identification.