Grey Wolf Optimization- and Particle Swarm Optimization-Based PD/I Controllers and DC/DC Buck Converters Designed for PEM Fuel Cell-Powered Quadrotor

The most important criterion in the design of unmanned air vehicles is to successfully complete the given task and consume minimum energy in the meantime. This paper presents a comparison of the performances of metaheuristic methods such as Particle Swarm Optimization (PSO) and Grey Wolf Optimization (GWO) to design controllers and DC/DC buck converters for optimizing the energy consumption and path following error of a PEM fuel cell-powered quadrotor system. Hence, the system consists of two PSO- and GWO-based optimizers. Optimizer I is used for determining the parameters of the PD controller, which is used for minimizing the route-tracking error. On the other hand, the I controller parameters and the values of the DC/DC buck converters’ components are determined by Optimizer II to minimize the voltage-tracking errors of the converters. Both optimizers work together in the system and try to minimize tracking errors while also minimizing power consumption by using suitable objective functions. Simulation results demonstrate the effectiveness of the PSO- and GWO-based design of the controllers and converters in enhancing energy efficiency and improving the quadrotor’s flight stability. For step inputs, the GWO-based optimized system shows better performance according to power consumption and the time domain criteria such as rise time and settling time. However, the PSO-based optimized system shows 24.707% better performance for overshoot. On the other hand, 10.8866% less power consumption is observed for the GWO-based optimized system. This power efficient performance of the GWO-based system increases to 18% for the complex route involving ramp and step inputs. Then, a 39 s route test was performed and the total power consumptions for the GWO-based optimized and PSO-based optimized systems were observed to be 168.0015 W/s and 179.9070 W/s, respectively. This means that GWO-based optimizers provide more energy-efficient performance for complex routes. On the other hand, it was determined that the tracking errors in the performance of the desired and actual values of both translational and rotational movement parameters and the forces and torques required for the quadrotor to follow this route were obtained at a maximum of 4% for systems optimized with both techniques. This shows that the full systems optimized with both GWO and PSO algorithms significantly increase their energy efficiency and provide maximum route-following performance.