Robust 3D Dynamic Region Reaching With Obstacle Avoidance For Quadrotor-type Drone
Basel Osama Raafat Ibrahim, Jawhar Ghommam, F. Mnif, Omar Mohamed Gad, Raouf Fareh · 2024
A novel control technique is presented for translational dynamics, incorporating backstepping, region-reaching strategies, and obstacle avoidance. Backstepping is employed to develop a stable control law that ensures convergence to a predefined region in three-dimensional space while avoiding obstacles. The proposed control strategy employs artificial potential functions to guide the quadrotor through a dynamic environment safely. A modified terminal sliding mode control approach is used in attitude dynamics. Integrating a Radial Basis Function (RBF) neural network enhances this technique by effectively handling uncertainties and disturbances. The revised terminal sliding mode controller guarantees the quadrotor's stable control over its attitude, even when faced with external disturbances and uncertainties in the model. The RBF neural network is employed to adjust the control inputs, allowing the quadrotor to achieve accurate attitude control, which is crucial for tasks such as aerial photography, surveillance, or any other application requiring stable positioning. The simulations are performed utilizing MATLAB/Simulink, which offers a virtual setting for testing and verifying the suggested control methods. The results indicate that the backstepping-based region reaching and modified terminal sliding mode control strategies are effective and robust.