Twisting Sliding Mode Control Design with Particle Swarm Optimization for Fixed Wing UAV
No Thumbnail Available
Date
2023-11
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Control of xed wing unmanned aerial vehicles (FW-UAVs) is challenging due to their
highly coupled, complex, nonlinear, and uncertain mathematical model, and underactuated
dynamics. To overcome this di culty a nonlinear robust twisting sliding mode
control (SMC) was designed for inertial positions, attitudes and airspeed control to track
the desired trajectories. Fuzzy based switching for inertial positions, and airspeed control
was applied to overcome the trade-o between chattering and robustness. Where as,
quasi-static saturation switching was used for attitude control. Particle swarm optimization
(PSO) was used for optimizing the gain parameters of the designed controllers.
Di erent set of trajectories; bow-tie, helical, and real data minimum snap polynomial
paths were prepared for guiding the FW-UAV. The required real data way-points were
taken manually using 'Google Earth Pro' as latitudes and longitudes in units of degree.
These data points converted into coordinates in earth-centered-earth- xed frame, and
then to the north-east down frame. Using these sample points minimum snap polynomial
trajectories developed through minimizing snap as a quadratic programming. The overall
mathematical model has been prepared in MATLAB Simulink for simulation.
Finally, the performance of controllers was evaluated against disturbances, model uncertainty,
parameter variation, and including actuator dynamics. Integral-time-absoluteerror
(ITAE) has been used as performance index and in all trajectories the controllers
were robust within 93% accuracy about the nominal value. The control was achieved
within practically accepted ranges of control e ort. Nominal de ection angles of aileron
(14◦), elevator (34◦), and rudder (22◦) were required, which are in the working ranges of
practical control, unity rad. The thrust force generated by the propeller has an average
value of 133 N, which is nearly equivalent to the weight of UAV, 132.5 N. The propeller
engine acceleration had maximum values of 10 m/s (376 revolutions per minute) and 30
m/s (1128 revolutions per minute) in helical and level ights respectively.
Description
Keywords
FW-UAV, coupled, twisting SMC, MATLAB Simulink, PSO, ITAE