Improving 3D Path Tracking of Unmanned Aerial Vehicles through Optimization of Compensated PD and PID Controllers

Nadia Samantha Zuñiga-Peña,Irving Barragan-Vite,Juan Carlos Seck-Tuoh-Mora,Joselito Medina-Marin,Norberto Hernández-Romero

doi:10.3390/app12010099

Nadia Samantha Zuñiga-Peña, Irving Barragan-Vite + Show 3 more

Open Access

https://doi.org/10.3390/app12010099

Copy DOI

Journal: Applied sciences	Publication Date: Dec 23, 2021
Citations: 6	License type: CC BY 4.0

Affiliation: Universidad Autónoma del Estado de Hidalgo

Abstract

The development of quadrotor unmanned aerial vehicles (QUAVs) is a growing field due to their wide range of applications. QUAVs are complex nonlinear systems with a chaotic nature that require a controller with extended dynamics. PD and PID controllers can be successfully applied when the parameters are accurate. However, this parameterization process is complicated and time-consuming; most of the time, parameters are chosen by trial and error without guaranteeing good performance. The originality of this work is to present a novel nonlinear mathematical model with aerodynamic moments and forces in the Newton–Euler formulation, and identify metaheuristic algorithms applied to parameter optimization of compensated PD and PID controls for tracking the trajectories of a QUAV. Eight metaheuristic algorithms (PSO, GWO, HGS, LSHADE, LSPACMA, MPA, SMA and WOA) are reported, and RMSE is used to measure each dynamic performance of the simulations. For the PD control, the best performance is obtained with the HGS algorithm with an RMSE = 0.037247252379126. For the PID control, the best performance is obtained with the HGS algorithm with an RMSE = 0.032594309723623. Trajectory tracking was successful for the QUAV by minimizing the error between the desired and actual dynamics.

Highlights

One of the most complex engineering problems is the limitation of algorithms to control unmanned aerial vehicles (UAVs)
In order for the quadrotor unmanned aerial vehicles (QUAVs) to follow a predetermined path in three-dimensional space, the control of trajectory tracking is proposed as Figure 3, where we introduce the desired positions xd, yd, zd to the Translational and Altitude Controller, which computes the control signal u1 and determines the desired angles φd, θd, which in turn are the references to the Attitude Controller
Calculations were performed in a Mac OS Big Sur (Intel core i9, RAM 24 GB, and 1 TB hard disk), using MATLAB 2015a (MathWorks) for each algorithm (PSO, Gray Wolf Optimization (GWO), Hunger Games Search (HGS), LSHADE, SPACMA, Marine Predators Algorithm (MPA), Slime Mould Algorithm (SMA), Whale Optimization Algorithm (WOA))

Summary

Introduction

One of the most complex engineering problems is the limitation of algorithms to control unmanned aerial vehicles (UAVs). The problem of tuning a PD or PID for a QUAV dynamic system results in a complex underactuated system, highly coupled in its variables, and many parameters to adjust, nonlinear, where it is not easy to apply conventional control techniques. Using these techniques, the dynamic model must meet convexity, differentiability, continuity, and linearity properties, leading to an excessive simplification of the model to obtain a global optimum, but failing to represent all the relevant features of the real system [15].

Quadrotor UAV

Strategy Control of Trajectory Tracking

PID Controller

Optimization of Compensated Controllers

Results and Discussion

PD Controller

Conclusions