Performance Comparison of the ANFIS based Quad-Copter Controller Algorithms

Abstract

Performing an accurate and smooth trajectory of a quad-copter is a crucial aspect in autonomous controls due to its non-linearity and under-actuated characteristic. Adaptive Neuro-Fuzzy Inference System (ANFIS) is well-known for nonlinear controls. This paper focuses on comparing the performance of ANFIS based quad-copter systems to identify the best optimization algorithm. Two famous algorithms called Genetic Algorithms (GA) and Particle Swarm Optimization (PSO) was used as the optimization algorithms and to tune the gains of the Fuzzy Inference Systems (FIS). The analysis was performed using two different simulations namely, altitude control and trajectory navigation. The final results were compared between traditional PID, conventional ANFIS, GA-ANFIS and PSO-ANFIS. PSO-ANFIS obtained the highest performance in our experiments.