Parallel Control System PD-SMCNN for Robust Autonomous Mini-Quadcopter
Kadek Dwi Wahyuadnyana, Purwadi Agus Darwito · 2022
A parallel control system with proportional-derivative (PD) and sliding mode control-neural network (SMCNN) has been designed to control the position and attitude of mini-quadcopter parrot mambo mini-drone (PMD). This PDSMCNN control scheme is implemented in the PMD system through a 3D environment using MATLAB-Simulink, which represented the real conditions. The PD controller is used to reach the minimum value of the gain for the PMD getting to take-off, while the SMC as a robust controller is adjusted using a backpropagation neural network (NN) to make the PMD more robust regarding an external disturbance. The software and hardware simulation have been conducted to validate the proposed controller with a mission plan as an input. Based on the simulation results, the PD controller shows an overshoot of 26.7% then becomes 0% of overshoot when the PD-SMCNN controller is implemented. From the hardware simulation results, it is found that the PD controller has 45.1% of robustness and the PD-SMCNN controller has 100% of robustness in the absence of external disturbance. Next, it is found that the PD controller has 26.9% of robustness and the PD-SMCNN has 74.8% of robustness in the presence of external disturbance. Based on these results, indicates that the proposed PD-SMCNN controller is superior to the PD controller in terms of robustness.