PSO–PID controller for quadcopter UAV: Index performance comparison

Abstract

A quadcopter is underactuated where there are 6° of motion with only four rotors to control all six motions. Varying the speed of the four rotors can produce thrust, roll, pitch and yaw torque which results in specific movements of the quadcopter. This paper presents the dynamic modeling of a quadcopter, which derived using Newton–Euler formalism and Proportional–integral–derivative (PID) controller for a quadcopter unmanned aerial vehicle. The PID controller is employed in this study due to its simplicity and easy to design. However, it is relatively difficult to determine the optimal tuning gains for PID controller which requires an ample of time with consideration of quadcopter dynamics and nonlinearities. There are several traditional methods for PID gains tuning such as manual tuning and Ziegler–Nichols (ZN-PID) methods, but both are time-wasting with unreliable results specifically for quadcopter system. This work proposes PID gains optimization using Particle Swarm Optimization (PSO) algorithm (PSO–PID) for quadcopter to reduce tuning effort with optimal results. This meta-heuristic algorithm is implemented to provide the optimal PID gains for altitude and attitude stabilization through setup iterations and populations with fixed boundaries. PSO performance is evaluated using several index performances which are IAE, ISE, ITAE and ITSE. The results obtained confirm that the PSO meta-heuristic algorithm works acceptably good with all index performances, especially ITSE, in identifying optimal PID gains for stabilization and trajectory tracking of a quadcopter. It is also proven that PSO–PID controller is better than ZN-PID controller in escape maneuvering of roll motion during wind disturbance occurrence.