Enhancing Trajectory Following in VTOL Cargo UAVs: Adaptive Control in Changing Payload Scenarios

This paper addresses the issue of controller performance degradation in cargo unmanned aerial vehicles (UAVs) due to varying payload amounts during a continuous flight. Linear controllers, such as PID, are often favored for nonaggressive UAV flights due to their simplicity and satisfactory performance compared to more complex alternatives. However, Vertical Takeoff and Landing (VTOL) cargo UAVs, while in operation, need to accommodate payloads of different masses, altering the flight dynamics. To maintain trajectories with an acceptable error margin, controller parameters, otherwise assumed constant, must be updated. In this study, we employ the metaheuristic Gray Wolf optimization technique to precompute PID parameter sets for a range of payloads in a 36-dimensional space, storing them in a look-up table (LUT). As the payload changes, we update cascade PID controller gains by interpolating LUT values. The proposed technique’s performance is demonstrated through simulations involving various trajectories and payloads.