Neural Network and Fuzzy-logic-based Self-tuning PID Control for Quadcopter Path Tracking

Abstract

The purpose of this research is to design adaptive control methods for addressing the stabilization and trajectory tracking problems in a quadcopter unmanned aerial vehicle (UAV). To accomplish these tasks, a comparative study of the Proportional Integral Derivative (PID) and PD controllers is performed. Intelligent algorithms (IAs) have been used to tune the conventional structure of PID/PD controllers. The proposed hybrid intelligent controllers consist of the neural network PID/PD (NNPID/PD) and the Optimized Fuzzy PID/PD based on the Particle Swarm Optimization (FPID/PD-PSO). Adaptive neural networks are deployed to schedule PID/PD gains, the improved back-propagation algorithm is used to update the weights of the neural network. Then, an effective control approach based on adaptive PID Fuzzy logic and Particle Swarm Optimization (PSO) algorithm has been applied. PSO algorithm is introduced to adjust the scaling factors for improving the convergence speed and production rate. Finally, in order to demonstrate the robustness of the proposed control methods, disturbances in the quadcopter system are added. The results so obtained demonstrate the effectiveness of the proposed control strategy.